Code_file.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "203190024_EE769_Assignment2_Complete.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PrRMqzbl-kED",
        "outputId": "6223a296-5394-4e8d-de42-6910dd08df86"
      },
      "source": [
        "pip install missingno"
      ],
      "execution_count": 262,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: missingno in /usr/local/lib/python3.7/dist-packages (0.4.2)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from missingno) (3.2.2)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from missingno) (1.19.5)\n",
            "Requirement already satisfied: seaborn in /usr/local/lib/python3.7/dist-packages (from missingno) (0.11.1)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from missingno) (1.4.1)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->missingno) (2.8.1)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->missingno) (0.10.0)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->missingno) (1.3.1)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->missingno) (2.4.7)\n",
            "Requirement already satisfied: pandas>=0.23 in /usr/local/lib/python3.7/dist-packages (from seaborn->missingno) (1.1.5)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib->missingno) (1.15.0)\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.23->seaborn->missingno) (2018.9)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bCFNXCjyufud"
      },
      "source": [
        "# Importing all files\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "import missingno as msno "
      ],
      "execution_count": 263,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qdKtpBeJ8qqA"
      },
      "source": [
        "dash = '-----------------------------------------------------------'"
      ],
      "execution_count": 265,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j6bc1YWH70HT",
        "outputId": "11330da9-b3d5-4fb6-a59f-f3ab7daf8fba"
      },
      "source": [
        "%rm -rf ML-Assignment-data\n",
        "!git clone https://github.com/py-here/ML-Assignment-data"
      ],
      "execution_count": 266,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'ML-Assignment-data'...\n",
            "remote: Enumerating objects: 9, done.\u001b[K\n",
            "remote: Counting objects: 100% (9/9), done.\u001b[K\n",
            "remote: Compressing objects: 100% (7/7), done.\u001b[K\n",
            "remote: Total 9 (delta 1), reused 0 (delta 0), pack-reused 0\u001b[K\n",
            "Unpacking objects: 100% (9/9), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l2LG_hHM77Yn"
      },
      "source": [
        "train_df = pd.read_csv(\"/content/ML-Assignment-data/Train.csv\")\n"
      ],
      "execution_count": 267,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GFqTEtYCDt0j"
      },
      "source": [
        "test_df = pd.read_csv(\"/content/ML-Assignment-data/test.csv\")"
      ],
      "execution_count": 392,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mgqndC772esO"
      },
      "source": [
        "# Data Analysis & Cleaning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2zVnW40Ea8gB"
      },
      "source": [
        "## Data Exploration"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "y3SmDJTD-0sJ",
        "outputId": "6c525292-c748-43cb-b10f-075ff3b99128"
      },
      "source": [
        "train_df.shape"
      ],
      "execution_count": 268,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(44550, 41)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 268
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Wol_MvM9JihM",
        "outputId": "4113a2fc-8a25-45de-ce7d-254725ca0cdf"
      },
      "source": [
        "df = train_df.append(test_df)\n",
        "df.shape"
      ],
      "execution_count": 269,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(59400, 42)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 269
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vcdLIUtS8VWo",
        "outputId": "859014b4-4532-4745-97fa-238820ca6565"
      },
      "source": [
        "train_df.info()"
      ],
      "execution_count": 270,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 44550 entries, 0 to 44549\n",
            "Data columns (total 41 columns):\n",
            " #   Column                 Non-Null Count  Dtype  \n",
            "---  ------                 --------------  -----  \n",
            " 0   id                     44550 non-null  int64  \n",
            " 1   amount_tsh             44550 non-null  float64\n",
            " 2   date_recorded          44550 non-null  object \n",
            " 3   funder                 41757 non-null  object \n",
            " 4   gps_height             44550 non-null  int64  \n",
            " 5   installer              41743 non-null  object \n",
            " 6   longitude              44550 non-null  float64\n",
            " 7   latitude               44550 non-null  float64\n",
            " 8   wpt_name               44550 non-null  object \n",
            " 9   num_private            44550 non-null  int64  \n",
            " 10  basin                  44550 non-null  object \n",
            " 11  subvillage             44263 non-null  object \n",
            " 12  region                 44550 non-null  object \n",
            " 13  region_code            44550 non-null  int64  \n",
            " 14  district_code          44550 non-null  int64  \n",
            " 15  lga                    44550 non-null  object \n",
            " 16  ward                   44550 non-null  object \n",
            " 17  population             44550 non-null  int64  \n",
            " 18  public_meeting         42059 non-null  object \n",
            " 19  recorded_by            44550 non-null  object \n",
            " 20  scheme_management      41718 non-null  object \n",
            " 21  scheme_name            23440 non-null  object \n",
            " 22  permit                 42214 non-null  object \n",
            " 23  construction_year      44550 non-null  int64  \n",
            " 24  extraction_type        44550 non-null  object \n",
            " 25  extraction_type_group  44550 non-null  object \n",
            " 26  extraction_type_class  44550 non-null  object \n",
            " 27  management             44550 non-null  object \n",
            " 28  management_group       44550 non-null  object \n",
            " 29  payment                44550 non-null  object \n",
            " 30  payment_type           44550 non-null  object \n",
            " 31  water_quality          44550 non-null  object \n",
            " 32  quality_group          44550 non-null  object \n",
            " 33  quantity               44550 non-null  object \n",
            " 34  quantity_group         44550 non-null  object \n",
            " 35  source                 44550 non-null  object \n",
            " 36  source_type            44550 non-null  object \n",
            " 37  source_class           44550 non-null  object \n",
            " 38  waterpoint_type        44550 non-null  object \n",
            " 39  waterpoint_type_group  44550 non-null  object \n",
            " 40  status_group           44550 non-null  object \n",
            "dtypes: float64(3), int64(7), object(31)\n",
            "memory usage: 13.9+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 997
        },
        "id": "goJ-80NL8XXX",
        "outputId": "2d81a881-9ded-42dd-bfc1-26f1033e4264"
      },
      "source": [
        "display(train_df.describe())\n",
        "print(dash)\n",
        "print(dash)\n",
        "print('Data Size:',train_df.shape)\n",
        "print('\\nData types:',train_df.dtypes)\n",
        "print(dash)\n",
        "display(train_df.head())"
      ],
      "execution_count": 271,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>gps_height</th>\n",
              "      <th>longitude</th>\n",
              "      <th>latitude</th>\n",
              "      <th>num_private</th>\n",
              "      <th>region_code</th>\n",
              "      <th>district_code</th>\n",
              "      <th>population</th>\n",
              "      <th>construction_year</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>44550.000000</td>\n",
              "      <td>44550.000000</td>\n",
              "      <td>44550.000000</td>\n",
              "      <td>44550.000000</td>\n",
              "      <td>4.455000e+04</td>\n",
              "      <td>44550.000000</td>\n",
              "      <td>44550.000000</td>\n",
              "      <td>44550.000000</td>\n",
              "      <td>44550.000000</td>\n",
              "      <td>44550.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>37192.623681</td>\n",
              "      <td>307.623224</td>\n",
              "      <td>665.490774</td>\n",
              "      <td>34.084084</td>\n",
              "      <td>-5.721980e+00</td>\n",
              "      <td>0.521549</td>\n",
              "      <td>15.338563</td>\n",
              "      <td>5.678967</td>\n",
              "      <td>179.112750</td>\n",
              "      <td>1295.782155</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>21399.827198</td>\n",
              "      <td>2402.742799</td>\n",
              "      <td>693.653179</td>\n",
              "      <td>6.555924</td>\n",
              "      <td>2.948828e+00</td>\n",
              "      <td>13.835868</td>\n",
              "      <td>17.672144</td>\n",
              "      <td>9.724205</td>\n",
              "      <td>478.995645</td>\n",
              "      <td>953.183961</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-90.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-1.164838e+01</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>18723.250000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>33.105585</td>\n",
              "      <td>-8.574241e+00</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>5.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>37183.500000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>363.000000</td>\n",
              "      <td>34.905600</td>\n",
              "      <td>-5.045611e+00</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>12.000000</td>\n",
              "      <td>3.000000</td>\n",
              "      <td>25.000000</td>\n",
              "      <td>1985.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>55647.500000</td>\n",
              "      <td>20.000000</td>\n",
              "      <td>1318.000000</td>\n",
              "      <td>37.176859</td>\n",
              "      <td>-3.327795e+00</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>17.000000</td>\n",
              "      <td>5.000000</td>\n",
              "      <td>210.000000</td>\n",
              "      <td>2004.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>74247.000000</td>\n",
              "      <td>170000.000000</td>\n",
              "      <td>2770.000000</td>\n",
              "      <td>40.345193</td>\n",
              "      <td>-2.000000e-08</td>\n",
              "      <td>1776.000000</td>\n",
              "      <td>99.000000</td>\n",
              "      <td>80.000000</td>\n",
              "      <td>30500.000000</td>\n",
              "      <td>2013.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                 id     amount_tsh  ...    population  construction_year\n",
              "count  44550.000000   44550.000000  ...  44550.000000       44550.000000\n",
              "mean   37192.623681     307.623224  ...    179.112750        1295.782155\n",
              "std    21399.827198    2402.742799  ...    478.995645         953.183961\n",
              "min        1.000000       0.000000  ...      0.000000           0.000000\n",
              "25%    18723.250000       0.000000  ...      0.000000           0.000000\n",
              "50%    37183.500000       0.000000  ...     25.000000        1985.000000\n",
              "75%    55647.500000      20.000000  ...    210.000000        2004.000000\n",
              "max    74247.000000  170000.000000  ...  30500.000000        2013.000000\n",
              "\n",
              "[8 rows x 10 columns]"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "-----------------------------------------------------------\n",
            "-----------------------------------------------------------\n",
            "Data Size: (44550, 41)\n",
            "\n",
            "Data types: id                         int64\n",
            "amount_tsh               float64\n",
            "date_recorded             object\n",
            "funder                    object\n",
            "gps_height                 int64\n",
            "                          ...   \n",
            "source_type               object\n",
            "source_class              object\n",
            "waterpoint_type           object\n",
            "waterpoint_type_group     object\n",
            "status_group              object\n",
            "Length: 41, dtype: object\n",
            "-----------------------------------------------------------\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>date_recorded</th>\n",
              "      <th>funder</th>\n",
              "      <th>gps_height</th>\n",
              "      <th>installer</th>\n",
              "      <th>longitude</th>\n",
              "      <th>latitude</th>\n",
              "      <th>wpt_name</th>\n",
              "      <th>num_private</th>\n",
              "      <th>basin</th>\n",
              "      <th>subvillage</th>\n",
              "      <th>region</th>\n",
              "      <th>region_code</th>\n",
              "      <th>district_code</th>\n",
              "      <th>lga</th>\n",
              "      <th>ward</th>\n",
              "      <th>population</th>\n",
              "      <th>public_meeting</th>\n",
              "      <th>recorded_by</th>\n",
              "      <th>scheme_management</th>\n",
              "      <th>scheme_name</th>\n",
              "      <th>permit</th>\n",
              "      <th>construction_year</th>\n",
              "      <th>extraction_type</th>\n",
              "      <th>extraction_type_group</th>\n",
              "      <th>extraction_type_class</th>\n",
              "      <th>management</th>\n",
              "      <th>management_group</th>\n",
              "      <th>payment</th>\n",
              "      <th>payment_type</th>\n",
              "      <th>water_quality</th>\n",
              "      <th>quality_group</th>\n",
              "      <th>quantity</th>\n",
              "      <th>quantity_group</th>\n",
              "      <th>source</th>\n",
              "      <th>source_type</th>\n",
              "      <th>source_class</th>\n",
              "      <th>waterpoint_type</th>\n",
              "      <th>waterpoint_type_group</th>\n",
              "      <th>status_group</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>7411</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2013-03-20</td>\n",
              "      <td>Germany Republi</td>\n",
              "      <td>1203</td>\n",
              "      <td>CES</td>\n",
              "      <td>37.228664</td>\n",
              "      <td>-3.256505</td>\n",
              "      <td>Kwa Abdalla</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>Mafurini</td>\n",
              "      <td>Kilimanjaro</td>\n",
              "      <td>3</td>\n",
              "      <td>5</td>\n",
              "      <td>Hai</td>\n",
              "      <td>Machame Kaskazini</td>\n",
              "      <td>90</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>Water Board</td>\n",
              "      <td>Machame water supply</td>\n",
              "      <td>True</td>\n",
              "      <td>2008</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>water board</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay per bucket</td>\n",
              "      <td>per bucket</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>enough</td>\n",
              "      <td>enough</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>18014</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2011-02-21</td>\n",
              "      <td>Cdtf</td>\n",
              "      <td>1409</td>\n",
              "      <td>CDTF</td>\n",
              "      <td>34.615481</td>\n",
              "      <td>-9.972601</td>\n",
              "      <td>none</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Nyasa</td>\n",
              "      <td>Ituni</td>\n",
              "      <td>Iringa</td>\n",
              "      <td>11</td>\n",
              "      <td>5</td>\n",
              "      <td>Ludewa</td>\n",
              "      <td>Luana</td>\n",
              "      <td>70</td>\n",
              "      <td>False</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>Private operator</td>\n",
              "      <td>Mgowel</td>\n",
              "      <td>False</td>\n",
              "      <td>1984</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>private operator</td>\n",
              "      <td>commercial</td>\n",
              "      <td>never pay</td>\n",
              "      <td>never pay</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>enough</td>\n",
              "      <td>enough</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>26103</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2013-01-04</td>\n",
              "      <td>Government Of Tanzania</td>\n",
              "      <td>1572</td>\n",
              "      <td>DWE</td>\n",
              "      <td>36.658778</td>\n",
              "      <td>-3.321529</td>\n",
              "      <td>Kwa Loishiye</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>Melebuki</td>\n",
              "      <td>Arusha</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>Arusha Rural</td>\n",
              "      <td>Kimnyaki</td>\n",
              "      <td>150</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>Saitero olosaita pipe line</td>\n",
              "      <td>True</td>\n",
              "      <td>2002</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>vwc</td>\n",
              "      <td>user-group</td>\n",
              "      <td>never pay</td>\n",
              "      <td>never pay</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>non functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>31140</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2011-04-03</td>\n",
              "      <td>District Council</td>\n",
              "      <td>0</td>\n",
              "      <td>District Council</td>\n",
              "      <td>33.628025</td>\n",
              "      <td>-8.860795</td>\n",
              "      <td>Mwanzechele</td>\n",
              "      <td>0</td>\n",
              "      <td>Rufiji</td>\n",
              "      <td>Darajani</td>\n",
              "      <td>Mbeya</td>\n",
              "      <td>12</td>\n",
              "      <td>2</td>\n",
              "      <td>Mbeya Rural</td>\n",
              "      <td>Inyala</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>NaN</td>\n",
              "      <td>True</td>\n",
              "      <td>0</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>wug</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay annually</td>\n",
              "      <td>annually</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>river</td>\n",
              "      <td>river/lake</td>\n",
              "      <td>surface</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>12286</td>\n",
              "      <td>300.0</td>\n",
              "      <td>2013-03-08</td>\n",
              "      <td>District Council</td>\n",
              "      <td>1283</td>\n",
              "      <td>District Council</td>\n",
              "      <td>37.655509</td>\n",
              "      <td>-3.669777</td>\n",
              "      <td>Kwa Kidaya</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>Mwira A</td>\n",
              "      <td>Kilimanjaro</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>Mwanga</td>\n",
              "      <td>Shinghatini</td>\n",
              "      <td>34</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>Churu water supply</td>\n",
              "      <td>False</td>\n",
              "      <td>2007</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>vwc</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay monthly</td>\n",
              "      <td>monthly</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      id  amount_tsh  ... waterpoint_type_group    status_group\n",
              "0   7411         7.0  ...    communal standpipe      functional\n",
              "1  18014         0.0  ...    communal standpipe      functional\n",
              "2  26103         0.0  ...    communal standpipe  non functional\n",
              "3  31140         0.0  ...    communal standpipe      functional\n",
              "4  12286       300.0  ...    communal standpipe      functional\n",
              "\n",
              "[5 rows x 41 columns]"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AN4Zb2xE8yiZ",
        "outputId": "52de8f2d-1a39-409b-b3a1-edd03463e0c6"
      },
      "source": [
        "for col in train_df.columns.values:\n",
        "  list_vals = pd.unique(train_df[col])\n",
        "  print('\\n' + col + ' is of type '+ str(train_df[col].dtypes) + ', has ' + str(len(list_vals)) + ' unique values, and ' + str(np.sum(pd.isnull(train_df[col]))) + ' null entries')\n",
        "  #print iff colloum has less than 10 unique values\n",
        "  if len(list_vals) < 10:\n",
        "    list_str = ''\n",
        "    for n_value in range(len(list_vals)):\n",
        "      list_str = list_str + str(list_vals[n_value]) + ', '\n",
        "    print(' These are: '+ list_str[0:len(list_str) - 2])"
      ],
      "execution_count": 272,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "id is of type int64, has 44550 unique values, and 0 null entries\n",
            "\n",
            "amount_tsh is of type float64, has 85 unique values, and 0 null entries\n",
            "\n",
            "date_recorded is of type object, has 346 unique values, and 0 null entries\n",
            "\n",
            "funder is of type object, has 1652 unique values, and 2793 null entries\n",
            "\n",
            "gps_height is of type int64, has 2396 unique values, and 0 null entries\n",
            "\n",
            "installer is of type object, has 1855 unique values, and 2807 null entries\n",
            "\n",
            "longitude is of type float64, has 43155 unique values, and 0 null entries\n",
            "\n",
            "latitude is of type float64, has 43155 unique values, and 0 null entries\n",
            "\n",
            "wpt_name is of type object, has 28991 unique values, and 0 null entries\n",
            "\n",
            "num_private is of type int64, has 59 unique values, and 0 null entries\n",
            "\n",
            "basin is of type object, has 9 unique values, and 0 null entries\n",
            " These are: Pangani, Lake Nyasa, Rufiji, Lake Tanganyika, Lake Victoria, Wami / Ruvu, Internal, Ruvuma / Southern Coast, Lake Rukwa\n",
            "\n",
            "subvillage is of type object, has 16618 unique values, and 287 null entries\n",
            "\n",
            "region is of type object, has 21 unique values, and 0 null entries\n",
            "\n",
            "region_code is of type int64, has 27 unique values, and 0 null entries\n",
            "\n",
            "district_code is of type int64, has 20 unique values, and 0 null entries\n",
            "\n",
            "lga is of type object, has 125 unique values, and 0 null entries\n",
            "\n",
            "ward is of type object, has 2080 unique values, and 0 null entries\n",
            "\n",
            "population is of type int64, has 956 unique values, and 0 null entries\n",
            "\n",
            "public_meeting is of type object, has 3 unique values, and 2491 null entries\n",
            " These are: True, False, nan\n",
            "\n",
            "recorded_by is of type object, has 1 unique values, and 0 null entries\n",
            " These are: GeoData Consultants Ltd\n",
            "\n",
            "scheme_management is of type object, has 13 unique values, and 2832 null entries\n",
            "\n",
            "scheme_name is of type object, has 2507 unique values, and 21110 null entries\n",
            "\n",
            "permit is of type object, has 3 unique values, and 2336 null entries\n",
            " These are: True, False, nan\n",
            "\n",
            "construction_year is of type int64, has 55 unique values, and 0 null entries\n",
            "\n",
            "extraction_type is of type object, has 18 unique values, and 0 null entries\n",
            "\n",
            "extraction_type_group is of type object, has 13 unique values, and 0 null entries\n",
            "\n",
            "extraction_type_class is of type object, has 7 unique values, and 0 null entries\n",
            " These are: gravity, handpump, motorpump, submersible, other, rope pump, wind-powered\n",
            "\n",
            "management is of type object, has 12 unique values, and 0 null entries\n",
            "\n",
            "management_group is of type object, has 5 unique values, and 0 null entries\n",
            " These are: user-group, commercial, parastatal, unknown, other\n",
            "\n",
            "payment is of type object, has 7 unique values, and 0 null entries\n",
            " These are: pay per bucket, never pay, pay annually, pay monthly, unknown, pay when scheme fails, other\n",
            "\n",
            "payment_type is of type object, has 7 unique values, and 0 null entries\n",
            " These are: per bucket, never pay, annually, monthly, unknown, on failure, other\n",
            "\n",
            "water_quality is of type object, has 8 unique values, and 0 null entries\n",
            " These are: soft, salty abandoned, unknown, salty, fluoride, milky, coloured, fluoride abandoned\n",
            "\n",
            "quality_group is of type object, has 6 unique values, and 0 null entries\n",
            " These are: good, salty, unknown, fluoride, milky, colored\n",
            "\n",
            "quantity is of type object, has 5 unique values, and 0 null entries\n",
            " These are: enough, insufficient, dry, unknown, seasonal\n",
            "\n",
            "quantity_group is of type object, has 5 unique values, and 0 null entries\n",
            " These are: enough, insufficient, dry, unknown, seasonal\n",
            "\n",
            "source is of type object, has 10 unique values, and 0 null entries\n",
            "\n",
            "source_type is of type object, has 7 unique values, and 0 null entries\n",
            " These are: spring, river/lake, shallow well, borehole, rainwater harvesting, dam, other\n",
            "\n",
            "source_class is of type object, has 3 unique values, and 0 null entries\n",
            " These are: groundwater, surface, unknown\n",
            "\n",
            "waterpoint_type is of type object, has 7 unique values, and 0 null entries\n",
            " These are: communal standpipe, hand pump, other, communal standpipe multiple, cattle trough, improved spring, dam\n",
            "\n",
            "waterpoint_type_group is of type object, has 6 unique values, and 0 null entries\n",
            " These are: communal standpipe, hand pump, other, cattle trough, improved spring, dam\n",
            "\n",
            "status_group is of type object, has 3 unique values, and 0 null entries\n",
            " These are: functional, non functional, functional needs repair\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hci_uObv-Mk8",
        "outputId": "124d10dd-057c-4654-d68f-c9507ff3e5ea"
      },
      "source": [
        "pd.set_option('display.max_rows', None)\n",
        "print('Number of null values:\\n',train_df.isnull().sum())\n",
        "pd.set_option('display.max_rows', 15)"
      ],
      "execution_count": 273,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Number of null values:\n",
            " id                           0\n",
            "amount_tsh                   0\n",
            "date_recorded                0\n",
            "funder                    2793\n",
            "gps_height                   0\n",
            "installer                 2807\n",
            "longitude                    0\n",
            "latitude                     0\n",
            "wpt_name                     0\n",
            "num_private                  0\n",
            "basin                        0\n",
            "subvillage                 287\n",
            "region                       0\n",
            "region_code                  0\n",
            "district_code                0\n",
            "lga                          0\n",
            "ward                         0\n",
            "population                   0\n",
            "public_meeting            2491\n",
            "recorded_by                  0\n",
            "scheme_management         2832\n",
            "scheme_name              21110\n",
            "permit                    2336\n",
            "construction_year            0\n",
            "extraction_type              0\n",
            "extraction_type_group        0\n",
            "extraction_type_class        0\n",
            "management                   0\n",
            "management_group             0\n",
            "payment                      0\n",
            "payment_type                 0\n",
            "water_quality                0\n",
            "quality_group                0\n",
            "quantity                     0\n",
            "quantity_group               0\n",
            "source                       0\n",
            "source_type                  0\n",
            "source_class                 0\n",
            "waterpoint_type              0\n",
            "waterpoint_type_group        0\n",
            "status_group                 0\n",
            "dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "vx3EqyB8-9LU",
        "outputId": "c313c1b0-8034-4f9d-82da-97c23c715b08"
      },
      "source": [
        "msno.matrix(train_df, labels=True, fontsize=8)"
      ],
      "execution_count": 274,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f95dc762510>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 274
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABboAAAKDCAYAAADcs7z/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydd5hkVfW13zVDTpIFAX9IEElmFFEkSBIUJSpJkiQRCSpIBomiCJJEguQoKiI5qShRBCQpCIjknEHy+v5Yp5g75fAB07e6upvzPg8PHWp6n6q6de456+y9tmxTqVQqlUqlUqlUKpVKpVKpVCqVynBlVL8HUKlUKpVKpVKpVCqVSqVSqVQqlcpAqEJ3pVKpVCqVSqVSqVQqlUqlUqlUhjVV6K5UKpVKpVKpVCqVSqVSqVQqlcqwpgrdlUqlUqlUKpVKpVKpVCqVSqVSGdZUobtSqVQqlUqlUqlUKpVKpVKpVCrDmip0VyqVSqVSqVQqlUqlUqlUKpVKZVhThe5KpVKpVCqVSqVSqVQqlUqlUqkMa6rQXalUKpVKpVKpVCqVSqVSqVQqlWFNFborlUqlUqlUKpVKpVKpVCqVSt+QNLrfY6gMf6rQXalUKpVKpVKpDCJ1I1epVCqVSqUyBkmjbL8maZSkfSWtKGmhfo+rMvyoQnelUqlUKpVKpTJISJodmE7ShJI+1OfhVEYokkY1vp6wn2OpVCqVSuWtsP26JAGHAaOBWYHlJX2hvyOrDDcm6PcAKpVKpVKpVCqVdxFzAgsBiwDnAv/s73AqIw1JoztZccBRwNOSbrB9XL/HVqlUKpVKk5LJ/Xr5dlVgbmAT23dIWhWYvW+DqwxLakZ3pVKpVCqVSqXSYxp2JZcBHwHeB1zavxFVRipF5BawL3ArcDCwrqT1ehWzaccjaaJexalUKpXKyKKTyS1pQeAmkgTwZUnvA54CVpQ0WbmvVSpvSRW6K5VKpVKpVCqVHtLwnZwA+DiwIfAzYAVJc0qaT9I8/R1lZYQxB7A+cJPtu4C1gK9Imq7tQJLU8FXdE/hg2zEqlUqlMrLoEq7nBfYApiVCN8ABwDbAT2y/YNuDPMTKMKUK3ZVKpVKpVCqVSo8oImDHd/JIspHbk2Rzv0YEyKOAyfs3yspIQtIOwITA14DdJS0MTAJMA7SeEdcQH04EXrR9s6SpJM3QdqxKpVKpjAw69w5JC9u+lVQhrQ/MBPwGuAr4M/CPvg2yMiypQnelUqlUKpVKpdIjGiLgtsALtpcBHgTWAX4LnACsb/u6Pg2xMvL4B7ne7gZ2INfYPsA2th9rK0iXXcmMwKPAbyUdD2wEbFsbYVYqlUqlSVez5PmANSUtZ/sq4OfAQaQR5XnAy6QaqVpiVd42VeiuVCqVSqVSqVRapikCFl4FnpU0j+0fk6bw3wbut33boA+wMmLolH9LOkDSB23/FjgV2Bm4C/g68F7g6ebjBxqzYVdyBPAe4BrgQ0SkuJZk5Y2o/WZToKlUKpXKO6PTeLJ4cq8NPA9cCXxM0nLARMA9JDHgn8DZwG9sv9y/UVeGG6o2N5VKpVKpVCqVSnsUEdBFFNsfuI5s3D4N/Be40PZtkj5g+9/9HGtl+CJptO3XGt+vDawHbGb7X5L2AuawvYakrwNrA6sCL7XldSrpEOA+2/tK6gjbkwBHAAfZPquNOEOBrs/1DsAzwEnAU833oVKpVCpvTjls/TXwOPASEbonBaYklls72r6kfyOsDHeq0F2pVCqVSqVSqfQASacRf8lJgQWI/+SXiBh4sO0X+ji8yjCmS3TdBXgEuAiYjfjA/5iI2j+yfVP5N1PYfm6Acd8Q10sp+RHAacDqQKcy4QbgGdtXdMY5kJhDiSLQ7E0qNABeAU4sDT8rlUqlMg46mdzl62mBQ2yvKWkuYuV2vu0rJU1j+8m+DrYyJGheM++UWnpVqVQqlUqlUqm0wDhsDa4FzgI+B5wB/Bs4HTitityV8aVLPD6SNJm8D/glcD+wO7AkcKztmxrX5fMDjDu6YVfyZWBOYo8yAfAjkpW3BHCZ7StgLI/6YUuX1csmwEeAI2zvDDwHbNCXgVUqlcowoNmUW9LJwCeB+yTNZ/sOcv9ascy1T/d1sJUhQZfFzZckTfVO/n0VuiuVSqVSqVQqlQHSla00T/nxdETg/hXwd+AyYErbd/dlkJVhTxGbXcTmTwGvA7sUi5DdgC1tXwhsa/tigM51OVDRuYjco4FLgE8AxwOL2/49sSvZjRFYqVBeb0lahRws3AKsJmkyUrExp6Sp+zrISqVSGYJ0HczuAtxb7lE3AxtK2p00Lz7fYbwyeCsji47IDVwMzGr7mXfy7yfozbAqlUqlUqlUKpV3B83ME2LjMKWky4BdgWmBKYBjgZ1t39y/kVaGO0VsFvBTcojyT+BgSRsBT5Brb0LbL7YVU9L3gIdtnwAsBpxs+0hJiwP7SJocOAXYoZSejwi7kq7n0Wke+15gR2JDdCgg4BjbT/VnlJVKpTI0eRPrifcUe5LjJS1IGhmfb/vKPgyxMsTo6j2yEHAjcIyknYGXgSttX/aWf2cErEEqlUqlUqlUKpW+0CVyrwnMaPsASacDf7D9c0kTArNVH9+BMxDPxuFMU3SVtC4Rlecp328LLA5MTDKqz2w59keAnwCH2f5t8ebeH/gdaa56DLCw7SfajDsUKJ/rJW1fImli4ATgGvL8fwQ8YXvffo6xUqlUhhpda6PDgAlsbyRpe3JweEj14q40aVwzo4BVSBP3bwIPAteTg+bHbJ/xVn+rWpdUKpVKpVKpVCrjQdN3klgarA/MUn69LvBFST+0/Qrx564MgMbrPUrSjyRN1+WfPCIpm7+OyL0QsQ65RtIeALb3A7YEvmH7zLZek2JTgu2/k2t7M0mr2H6ZNGH8IClF32AkidxdXvuzAPtJWsP2S8DXgfWADYF9gAUkrT8Of/5KpVJ519JYG50B3A5MJ+lA2/sAEwJblSSASgUY65o5F/iA7auBzW3vAvwN2Bx49O38rZrRXalUKpVKpVKpDABJOwBTAXsBBwB/sn1C2cTNb/uGvg5whCHp+8Dctjfu91h6TSeTu2z+9gZmBo4DbiPi81S2t+th3FFEqPgBycI7qIzjPmAZ4Fbbl7Ydv190Pe+vkgOqR4CjyWHWXUTkPs32nyXNAIy2/VDfBl2pVCpDEEmzAHvY3qB8fxLwlO3NJc1i+/7+jrAyFOjqcTM/sI3tDSXtD9wLPEz6gpzf6T3yVtST50qlUqlUKpVK5R3QyXQtX09Dsj6fAV4EfggsI2kj269UkXvgNLNlJc0GfAB4SNJ7+zeqwaHhEX0c8AJwA/ApYEHi+/6qpA/2MO6BwM22b7d9K8kc3wuY1/YhI1HkLt9eAnyWeHEvCWwBfJ747p9VRO7Rth+tInelUqmMk8eBWUufB4CDgU+UzO4qclc6ntyvl4bPnyMZ25NJ2oY0cDfphbHL2xW5oTajrFQqlUqlUqlU3jYl8+S1Ir5+D7iQlFl+EvgKcBawPTB1/0Y5cujy+VwMuAPYCdgGWFTS5bYf7OsgB4cniMD6d0nfAL5MGjPt0WbjyXHwLHBF4/t/AKsyxqJnRNCVyT0dcK3t75dGmycAL9veUtJktl+ANAbt55grlUplqFLm1BclrQ78qhxMLwpsAKwnaeraxLfSWE+fBDwF/Ac4xfZZkuYBdiCN3F94J3+3ZnRXKpVKpVKpVCpvky7fyQnJpm1B4HKSabs6cL/tm/s3ypFDozHRH4hVxlnAPMDPgIWBJUtzxBFJI5v9b8AiJaP9duD9pAFlTxKXGnFvBD4g6WOSPgZcCbxq+5pexO0XDXuY3YHlgQ9LWs7288ARwFwl8+wdbbYrlUplMOjuEzCY/SvGFbvMqaOLmL0y8GvS42AGYCmSpVupQA7PbwW2BtYBnizWf2sCO9m+8J3+wSp0VyqVSqVSqVQqb0HTrgSYngiAPwE+DFxfSirPAK5p2B9UxpOujfOXgFNt7wC8Tl7n+Ykf+l9Lc8QRSce3kgj9o4l1yAHAbsCswKd7HPcv5DVfg9h47GL74V7E7AddYtBRwCS2jyNNNncqfvAHA1fXDO5KpTIU6bJ/+FjJlh6Udcj/L3bJ1p3A9rO2ryL647LAurafHIzxVYYe42jefAc5+DgF2JH4cm8N7Gf7ovGKUdfhlUqlUqlUKpXKm9NlnzG37dsl/RKYHTgMuAb4JbDOu8RGo6eUjfNr5ev3AK8R//Ofk43Q9ESIXNj2c72KPdiMK3ZHiC3ZcZMRu5IZgTmAPYlg8J8ex53Y9kuSZrN9b5eX9bClq9HnFMAKZJP9Jdv/kTQvMDFA9dqvVCpDmSIeXkCqfxYGvmf72n7G7moyuBw5nP2V7acHY1yVoUdjPT2KHJ7fR9Z1+5JD9aOBPYCDbP9+fONUj+5KpVKpVCqVytuiuWl5N9FYlB8ErC1pCeBQ0njSpFHgj6vI3Q4Nz8ajgeeB64CriF/0DMBawGY9ELmb/usbA9cDD9u+u8044xO7/PzFci0+Tsp7t2hB5H6ruAJeKQ9/dCCxhhJdIvd2xGN/A/JcfyFpU9v/6OsgK5XxpHkYNVIOpipvyVLABbZ/Iuk2YKp+x26I3KsC3yEHs1XkfhfTSBrZlzTYnptUjP0I+By5F493JneHmtFdGVK8WzfQlUqlUqkMdbqymo8CrrR9VL/HNVhIOpGIgHcD3wU2AR4i2bWTjzTP4n7QJc7sQ4TVP5Fs+a8AHyf+3H8bH8/GtzsG0nzwTuAZ4P+AHwyGN/PbiS1pHWABYPu21sxvM+5aJFNvy5G0Vi+f63+Wb2cnovfSwKbAcsBLVSSsDCe6smiryD1C6a7EkbQwmb9EKs3uJE2bt27b3uvtxra9uaQVgO8Dm9i+rc1xVIYPXfPSasBqwN62b5C0O6naO6ytg5Dq0V0ZEkj6pKT5GxlTlUqlUqlUhhANkfskklm7k6Q1+zysnjGO9ci/gdNt/5qUVZ4PvN/2LVXkboeGyD0DsYO5g2yedyElrY+RzPlWRe4un+aPA48DBxLB89zykMnajPkOY09eHrsyyWY/fqBi83jEXQc4fISJ3HMCU9re0/ae5CBrX+Bs4Mu2XxwJImHzve563ysjjOaBtKQjgCMkraA0sa2MEDpCs6RRkraTtDmxDbmJNCq+mxwQn90rkfttxD6n/JOpgA2qyP3uRWP7uM9LmrdfC3xO0syk58jkJHGkFaqgWBkqzAocLWmeKnZXKpVKpTJ06BJG5ib2CVuRhkI7SPpmf0bWO7rEgg9ImgK4AThS0lTA7cAtRESYtwfxm8LU6P/fY0cakr4KLA48DGxPbEtuA04GJrb9asvxRnWJmf8E7gcuJpv1G4hNzbRtxn2HsaeStDjxj/6O7VuGY9yhhu07gb9K2r786DZgQmBnYIKRIAqP472u1qUjmMaB9A+BB4AjgQ8RQUkj4ZquvGHvJdKnYYry38nkEH4/4PPAXrbP62Psc8s/Oc32XW2PozJ8aNijnUiaTK5Jsv4nB9YG3md7J9v/aitmvdFV+kqnnMr2mZLmAg6UtJXt26qNSaVSqVQq/aWr1HB6YtXxfkkfLeWG+wI/kPSC7ZP7OtgWaRy6XwJcTTwDvwZMRHy6PwisCqwOzAS06uXbyGxeDfiQpAuB69vOzBqi3Es20N8A9iIizQ7AHgP1bOymrEM7wtAxRFy/hZTQXgS8l/ivH2z7vj7FPsj2g5L+C6xk+57hGLcrfl+ypJuxG1+fBHxd0q+J//uewGeBV0ZCNndjLjuNZNBNIumIXvcT6Of7PJTGMFh0PdePAvMB59m+RtJTpDLmLNvP922QlQHT9T6vRdYlc5V+A98FDrD97aESu3O/6cV4KkOfLj1vLbJWPhS4ElgF+AvpjTFh67HfJXN/ZYhTssEWAp4i5ZNb2r61it2VSqVSqfSHRnmqSJbn1CRb53XgJ8DpROzdj2x4Nrf9bL/G2wZdQti3AWwfIul64GXg67b/XTK8FyS+xkvb/ncPxrImsB6wf/lvS9uXtB1nqCBpGyJyXwHMBcxn++fldxP1oPy6+V7/jFh3XEDe1ynIIcfMwJO2/9rP2OryQx1ucUu89wEP9WNd/3ZiS1qE2BO9Hzgc+KLthwZpiD2h835L+jFwD3l/DyXeudcP0hgGbS/3ZrHeZYL3TsR+ZzFgSWIB9XD52Tdt39HH4VUGSOMzvT1Zex0OjLa9gaQFiCf3pr04FO9n7MrwQ9J8xObwAZLlL+Aj5CD5BGJ1swRwai8O4Ko9RKUvdJXkTkrE7QNsb0c20wdImq+K3JVKpVKpDD6S5miI3NsB/yINGPcF/ksyM/4JfIscUk8PtCKI9YtmiX+xCzkOOLZken4XOAP4peLVbGK7tnxbIvc4SspfLHGnI97Bf5D0/jZiDVFuA14Cjia2OAs3ftf6tdV4r5cF3gM8bvtq4A/A9LZvsn1h2yL3+MRuS2zuV9xybc9FrH8ulLSIBsmm8K1id762fQUwMbAusMZwFrkbz6kj7t4CvErm732ARyQt1av3oPN3JW1LmtD1HI2xmxol6XBJW0j6YjmsGbEi9zjuG1MSu5I/lv++AmwO7GP7jnE8vjIMKGuS5md6OtLccSNSpfEHYEvgxB4cCvctdmVYMyGxs/kTqYK8DViUCN/3AUcAD/eqyqQK3ZVBpyubZHLb/wX+TspSsX0K8DSwfP9GWakMX+oitlKpDATFSmy58u37SVbYg47H4jZkEz2fYyPxDBG7d7D9QguxmwLUoM1lXSLJocAvgBlsP0cE56eJhcb2tl8oC/Nf2769rTE01kYzlSSA54itwvK21yFZ3Zu0FW+o0HmfbZ9j+yxyPd1FLFv2K7/rySGKpBVJpcLJwHSS1iBe3EtKmrOX12C/Yg923Ma6/2rSaGom4EYPQk+et4pdft+xZlqJZJvtYPufvRxXL+k8J4WVJL2XzF+fJU3i7gB+BbzQdkJR47Pc+bvHAvNKWqjNOOPCY+x4DiIWW/eT9/NrvY7dTxr3jY0lTVeSxv4F/J4I3SeRSoUnJE0ykkX/kUxJPBgtad3yo8NIdR221yR9FSayfelIiT1Yh6GVdmkcjPydWJNMC/zV9v3kQOR2YCnSWPzsno2jznWVwaZR9vJd0tRqYuKH9yXSfOh54FrbB/ZvlJXK8EEtlhZXKm9G1yFlveZGKJIms/1CWajuTXxdXwK2BY60/RdJHyObmqtLdvMEtp9pcQx9KTNvZK+PBm4lguvqxJ7l88Apts9te3xdn62NgS+T7MsNydroC8AjwKTukffmYDKu+aOxNmy+FpMQi4W9eziWVcmhzQ9LlvPqxKf5INsXq4e2C/2KPZhxu97PKYBFgMlIMsuWACXhpXXeRmx1DuckrUyqJ9Zzi82w+kWZy44i2b1/I8ltjxFxamngl7YvbDnm1LafKl8fQhqb3khscR6zfXkvrmmN3UdiWeA7wCG2z5P0WWBR2/u2GXMo0D2PStqb3DcOtP2EpN8A/7W9lqSNyOHWoW0cSL9b6ZpTBmWdIukjpAnzNUqj5m8S8fAFsj44wvYZ5bGnAPfZbqWKos+xO2uCUaQS5QrgAdvXtPH3K71BYywPR5O1xd+B2Ygf9+EkOWZ22xf3eiy1GWVl0JD0IeKP95SkDwOL2/6y0mBpE9vbSroRmK4xab5rPNUqlfGhfEY6nYx3JVk7vwXuqULkyKJf4nJnHm4s7lcDZpb0V9tXDvZ4Kr1D0qzANyT9EpiXCCQ/JWXPPwV2VLySLy2Pf0MoapmNJc1ie5ce/O2x6BJeliGi8o+K4DcF8GvSHf6Xnc9fD0Xu+YkotBbZRB5HGjKeDszWEeCG89qovN4dS5zPA3+z/dybPJ//A5aWdFhHQBtg7BlsP1q+XhqY3PYZkn5WxN/fkFLbBYDRbc65/Yrd5+f8xuEF+Rw9QeyOfg9MQrKKJ5e0nu3/tBFzPGKvTgTwzYD1h7PI3TUvbEL2U1+V9EFSoXOT7ask/c72Yy3HngRYXdLfSeb4HcAsJKN6KeBhSXe6B3YwDZF7aWLRcirwUUnPAK8AXyz3tEeH67w5Lhpr/52ID+4vgE2BLSXdScTII8tjj5Q01XAWuXt56Ph2adyrv0Ky5G8chPvxPcBpkj4A7Gr7S5I+TXopTAJ8SdLfbP/b9hqSZhoJsRvz92FEHJ0Z+Ei5jnsuklbGj8a8dDZwFbAGqSq5BFif2IjtORhjqUJ3ZVAopzpLAx9WMrmfJoueLcnCc1el6dKZjeyKYbuRq1QGg67PyBnkJjId8HPSwOyh+jkaGXQdaPyMlKU+6lg99ZrJiYUCktYCNgZ+CPxW0sqOr2llBGD7PknPkuvrBNvfkvQ4cDCwFfF2fb3x+F7NLecBX+3R334Djd1sczrbFygZ6qtK+pftE8rvZuuFQANjbZy/SdZJswKv2T65jOVSYJmRIHLDWPYCR5Msn4skXeuukmfFTmNKYOWWRO7pgMUkvUwONK4B5pP0eeJ/Pn8RgC8hc97cwOWUuW84xu7nc4ax5oefk6yun5KsrlWBU4hHJ22L3O8gtmw/Wu6rm3mYN+lrzCVrEbulhyQtZ/t8SasAnyHCwxNtxpW0BTmUe4LM3ec4VktImoiIVB8HVpF0GHndByxYdh0STk0qYH5H7DqWIAe07wF2sf1ImXeGPV2C7+nEBuAFYG1iR/RBcnB1ou0/de5zbrHqarAp73XH6mhL8h4/Y/vOPoxlfZLZvEL50YSkUXUvYsn2k5IOBvai9Kpw+ipQDjS+Tq5zyu9aWav0K3bXGmcVcl/azPbtZR6bfaAxKu0j6QfApU7G/ZzAn23vLekq4Kmyvj4HmNL2g4Mxpup78y6keaMfjJt+R6CxfTDwILA9cC/wJLAjsDURTT7ePGkezhs5qD7Jld4i6b2d0+4ihvwROJ5kA/68PGay4f45qiTLtvE+HkpEyOuA3SW9r4dxJWkG4GxJnyw/HgV8myxsLwGukTRbr8ZQGRzKYXSHy0nJ++ySprd9CfGGPhy43fHc60lsSd+U9G0iQi0s6UttxuqmcXh0MbCdUup9Hrm2d5c0p+3j3YNmhE0kLQN8mlilXAocXMSMo4AtOxm5ZcwjYU7/ATCZ7aWJSPNhScvBG1lcK5EDtT/bfrKlmM+U//YGZizv6w+ICDor8D1JSzs2GmcCxzr+7MM5dl/i6n999h8i1/erJLv6cdLw8UrbfxlovAHE/jOA7d8OZ5G7a183B6mEOA04F1hKsbRYDOg0GG0tK1bSzMTS6RlyXW0MPKfYW1HiHUAO6N/j0Er8hsj9RZK5/R2STb48mccvIPezp5qPH85oTC+J0WX9dyWwr+2fEHurb9r+te3dywGHPAKqOxvZvScC05CKpw0kTdXr2F2fr4mBD5PrfClJu5AM+sl7Ebtzzdr+PfBFYF1Jqyu9FFayfRO5vtdtW3foV+zG3vbDwE0kM/hLSsXhU8BXJE1edZYhx7+AH0v6OLnvLiLpYmB34F+SzgUmGSyRG6rQ/a5A0kJl04Di0bW9pA1grBtHz2gsRDYg2TvvJV5LPyKb5x8Sr8/vlccN64mrc9MdCQuqytCkLKg2VcqO9wQ+QRZe15LyoOvIpnW+vg2y0gpKieA2ZcEHpYcBKUveDnhVUq/eZxWBbU9gT0lzk+ZOFwBr2F6LzOMb9yh+ZRDobISVxmzfBz5G7stHAccrnqcvASu5B53RPabJ0eRElFiQZOY9CKwmaZq2Y3bxXZIFuC2xazmTCEQXAZP2IuA4Eg4WI/1KXrO9M+lIf3Z5by7r/jfDja6DFMjm9VXFauD35PkuWDavqwLfJ5Z2/2hrDLZfIQco5wI3SVqh/Pw423sSa4dFJE1g+6UWRe6+xe5H3HLNdho9bkWyHs8gmad7AwJOIH7/rQpw/YzdT8peblQRnf9NKvs2cRq7/hmYCjirB4cK8xFrjAWKwH42EdOPJ+LfusDPJU1J7iGfkzTdQOeyrsPRaUm1wldJVvM2xL5jWbJWeQz4rNLcd9jjMRUxOwArE9F36/Lra0lz2Wkajx/We9Gua+VDZA16AGmQ/SdyH+npe9vQMdYtY/gHsWOYHTiLZM+37pLQdWg3ymnk923gK8B+jKnM+A/wszbf637EHsd7/UPSO+I8Ukm4P1mv7Wf7+eF+bY8UNKbx5K/JtXEAaT55CEmKegE4BjjYLVtmveXY6jUy8pG0ILnAziLlTEeQcq7r3cPmHGXB2bk5zEHE7e8QH7Fvk4XXHrZfHNe/GY4omY0fIgurFckC8MHh/JwqQ4vOZ0TSAmQRf7PtZSVNSDYXV5LF7z62L+jnWCsDp2wQlycL6cPJocYhwIHAL8nBxi62/9bDMSxBxOxpSSbcYqRE9p+kBG3TXsV+t6Ah0NxT0onkkOw50pBtEcVOYw1gT9t/aDneVsCEtn+sZFLfCTxve7cioCwGbFDGcktb64Pu11rJinuWlHwfSsS/9wFf68W9u2tt9AngSdt3KWWfE5Eu9P+VtKpLv5KRQNk4702EsGdJqfciwHUl8/C9th+WtCjwiO3bejSOSUmG7xJlDNOQQ45FgS3IId4rIyn2YMctgsUJxC/5aeJ7vym5f8wFbN6r+a6fsQcbSTOSe/CdJZFpBmJZcijwY2A3IjAvB8wDXO1U6LQ5hiWJVd4ewPTkM74+scJZmny+L5M0PWMOzwcSr5PRPIoIvFeR/eQnicj/NHmvf+L4kc8AvGz76YHE7TfN+5akHwFz2l61XO9HkGzuLwI72T63j0Ntje57fkkk25rMY0cAfyCVAlvZfqDHY1mdaCcr2n66HA6+KuknpFJho5bifBiYwfYlkhYHrvEYS9fOtT8FyYxt22e/b7G7xvEZ21dKWojsPU4l2cIrk+SDo20/0qv4lbdP47oYDcxBPN0/QgTvdcjcPD1J4mj1oPVtja/qbyOXzslYEcXmIKLI+bb3LYve44A7bW/f43F8lSy8vgrcaPuwItLtQ/zDTuuMd7gLwkVs3JySIWb72+Xnw/65VUI/30uN8ZPt3FiWJou+Y2z/qgg2oyBeu/0YY9t0iUJ9b0TTDyT9nCwW7gbOJ5UxiwEzAUc5ZYW9iv0J4CekId6SpCnfuuSEfvgQduIAACAASURBVBaXjMs6x40/jc+zgEVdMngHK275eiZiK7Y38bI9k2Q0P082Nc/2IP50wLFE3L3Q9v6SjiAL4s3KY75OStG/a3vAHpiNg8JO897niSj0KjlI+iERAE9wj73nFV/bVYEbgFdsf0/SjmRTsG1HeBzOny1J07hYj0g6HniYNKmbmljYPc6Y5p+Ptxy7ee8Yl2DyISKAnl6E9plIdeE9wzX2EHrOHyJr4d2IIHUMcBkRP0e3vT7pZ+x+UYSFeYiY8Dliy/JzIjC8RMTAfWz/prz3SxAroAH7c4/jsHApsi7Ym2TwHQJsYPvG8vtW1m5d7/NvgbvKr24ha5JJScbpYbYvHM5zZ5OuNcJnyBx6DHCq00diAtIYcOoRdH03m8keTvyhfw0sRGyXribX/s9sn9Or+OXrCUmy3o3ECmpx4DbyPnze9k7d/2YAcSck668FyOf38K7XolO1MjdwrxvJggOlX7HHMX9/CzjP9nmSPkISejYjhzkrkPfg+DbWhJWBU66NU8h6+v9Ic9xXSHPcjW1f16+xVeuSEYwLZaM4EcmmXlXSao4f37qkBKhnlBPBHWyfT077Z5O0ZxnLBR2RuzPeXo6ll2hM2cYrpMHSBcCtkuYvD6mNX0cAzRu+pE0kza9kqQxW7I6f7NGStim/2gDYTNJ2pFnI6yNooTu6OS80RLlhW8L/duk8R0m7kezH7YiX7RdI1v42wKZti9zjeG1fB160fa/t40ip6DnAFFXkbofGBvYo4gv9A0nz9vI6L5+t18tcNhXpmXE78Vg9i2RKHQvM0yORe3QRNtcmwsgMALY3Bl5XSvAhWTwzk2ZPbcTsXKe7NP7myWQTfQ8pjT27FyJ38/2UNCdp1rMUyWRfSdJhtvcC/uRGdu1w/Wwpljdzl69FhKif2/4FY6pBLiBlyG2L3KO67h1viL/l+2echkkbFsF3lO2HWhJ8+xK7z89ZXbH/SayHriBe0X8lIuxkvRS5Bzt2PylC8wPA+8l9+irHfmZLcnh3IrCj0uPjGWJd0obIPcpjrK52VLIu/0EOFbYj65UtgMkaY21D5B5V1t+jyjr4AXIouTu5f9j2McBqti8scYfl3NmkXN+dNcJ3iMD9WZI1/3VJa9t+1fZzI+z67rx3e5LP86nEjvE2kqh3N6lm7LXIPQe5lvcn9+xpyedrRVIZ34rIPQ4d4UGgmbA4UZF1Xpe0GhETW7Fs6XPs5r56HZIYeQXwMcV7fzJij/K8U+X1e+DXVeTuP4017fHAY7Y3BL5HkqMeJevsifo0PKBmdI94JG1MStbWd0pt/o9MYifaPrbxuLZKgrszSFYi2SM/tn1x2VBvQDK2Dm4zdr9Qsmr/TTIL9i9fH0cyHyckZXTzAHu7Rc/HSn8oE/sZwK3EB/EO4vfYuhj0JvF/QW76fyRZMzsS4XN1YmPS0yzEwaKx+BlFBMAbSTbgOR7GnePfCv1vptRKwKfIZm4i4n32KLC7c2DZq3EsSiwsbidi5KdJlty2wMO2D+9V7Hcj5aDqg7Y3lLQ5MJqIzTf36v5YPlsXksOLz5CMpWXJxmIBkil1dg/idj7bnQqVGUg2yKVkM3sG8FXb9ygZRjO3JAJ2yt23JBZA37L9jKSdgI8Tixa5xQypRvzmxnla208oTa2WJtVumxJBbufOaz4C1kaT2X5B0heIT/SiwH9JVcqHSTXS191oQt5S3Gb242kkoeMiks36rMa2APgccG1b73m/Yvf5OTcz/s4ih1N/IhYpi5Hs4kXowXzSz9hDBcXm6b0kY/scIgA+ZfslSTsDZzpN49qMOYr0NfgHSZp7lMzbs5L+IRu5B/0cyvt8AvEC/zA5FLxA0vLAaiQT9KU2hPWhhqTDyXrsr6SZ7zFk7X88sI57bN3RD5R+DSuTxLjjlOzeTYGte3GfHkf8rcma6InGGEYBhxENeLOW4nRsUEYBqxBR91xJp5H57HfAvEVHWR3YCNiiHOoN29iNMYhk6z9G5u6rSM+SKYGvkXXRxW3FqwyMcexTNyQVVEvb/qfSUP422xf1a4wdapbpCENpQDGj7dvKCd0HyenXMpI+TxpVdDw336CtzVRjI/dFMmHdRU52Fi4L0ouIt2xnvMN9IzcpOVmfj5z0PgbMQgShk8rvPkZukCNC5O6e4N4NdF2nHwRuJsLjX4DfkBKdwRjHe8kC/w/Ar0iDpZ2BQ2wfMRhjGCwar/eRxDf4XrKBuo6UrY041MiUImV615FSsH+U728i2a8H9ljk3pD0UTidHN59iwivRwGP2969PG5Yz9/9ZBzz6JPAc0q38qPIa74suYe2LhgU1iPZywdKup6IvDtKej/xzr6zzWBKhu/Vtp8q185r5XV4VNLXyPpkJbKJvaf87hWSaT0gGiLgHkDnmv0RsJntPZXKifndoxLLxtpoQ2AFSXfa/r7id3kvyXw5vinGDdfPVhHengIekrQ2OUT4KxHD5iDPdSFg17ZFbhgre/Qg4HrS4PRTwGSSLnLx6pW0CmnotiItVTf2K3a/4nbdA75A7AQuIwdJj5BM6qmBM3ogtvYt9v9nHING5x5i+1ZJ9wJTkDl9cWBNSbeTbO9WkjC6nufswB8du6kryUEWxPZg8zZFbo1te7I78KjtUyX9h2Q0L0buldv2cl3UTyRNQtZgVzrexY+QdcKztr/Q39H1BiXJ4yFysPFZpZHuo8R+aRZSCdV2zE8CH7V9lOJXvQDF9xzYTbF6+inwkO3dyr8Z8Oe/CM0i6+3JgQkUi9dvkEqJlYCdlEqwlWlRaO5X7K5DyvcA/7W9cYmzDnCR7cslHWv7qYHGq7RHY5+6GzkMOYEkop0paW9iybdr/0Y4hip0jzwmBTZUyn7vIiL3QmQS2Y2Icmfa3qdXA5C0HpkgryHZzPeTBehXJD1o++bOY4frRg7iZeU0jDoAWIssqE+yfbWk7YENScbt6bZf7edY26IhTowiG6oriPA1YpseNhcxSgbew6RR2TnkEOc24GRJm3iATXbGFZsxnmjbk6Zlh5cF3x9t/6gcbt3dZtx+Mg4B8Hriv7gVcDDwtKT5bN/alwH2kIYYdyzJajBp6vEXssDfjJT+91IwAJiflP/eUTbQe9heTdIpHuO3W0Xu8aRrHv0BOTi7gWS7foZkyB1MPNDbFAy6P1tPAEtIOgv4PvCfkonx87YPMyXNQjLw5pd0hO3nusTuxyV9DJjd9g2d37U5BnJ4MA+x/HlM0mGSfmF7k86mtZco1V+fJYdIu0nahfQq2ZVkIR5QHjfcP1uTkfLVuYhH449JgsVtpBLpTpKl32qzSY2dtTw76WvwmNMI7yUihk4AnFoE+DVJE8Y2BN++xO7zc24KFWeQ7Ok7SuyXS2zbPnmgsYZY7OnIQeBDkqbwICWx/P/iOln7F5L7yJG2ry//5ltuqcloYx28ErFs2F9pxPcL8tneDDits0ZoA41dqTARWQ+uImn+IvjeQwTg02z/va24Qw3bL0o6BdimCEkvkNdiI0lP2P5rf0c4cMZx35sCWN72RpImJ5VPkwL7Os1Xe3GffAjYQNLLto+X9H0yZ85GDlPOJ97Ru5UxD8h7XrEBud+pxN2dWMUtX353DLC97XUlzWz7wbLf26KNvWafYzfn75NIFfz9khawfXPZd6yoHKKNyMSm4UjXHmJvUsHzKDlc3ZckMexBMvD/pFIt0JfBFqpH9wjDKV36DynVu9v2n2z/hJTZ7EOaQbbaebpMVE0WAfa0/QNyI56WZGod1RS5hzPlw/6K0gDkc7YPIyVkS0qarRwkvEoag4wUkbvp/3g8seyYDFhfyQAccTRuxqMkHUa6e0MWO68SYeo40gG6VZEb3vDZ7yyiTgQWVzI+/w2sI+lqsugaKXYlHSsDSVpb0rTkdd6JiL23AL8lWQcjhiJ4dpiVeNd+s8wrfwSwfSKwnlsu3+s6yFmqHOY8Rbw2O3EfkzRJFbkHTln4dV6708gh9KeAT5KKoP8SO4tJbf+nxbhviOuSvqlk7FxB3uspiPh4MnBLDwRmirB2NtkwrleEms5G5/XyujwF3KccIvfi+jqdZB6uXcb0LWBCJQO5p5TP1SrE7/4Bx4v8A6Th00629yiPG7afrbIewva15NqaEfi77QdJZc5HSIXbC70Sucu94/Pkc3QY8Lyk9W1fTSqhrinz7dLA92zfPlxj9/M5wxverSJC55XkQGO0pHVsXwVcQpJcWqdfscs1Pg+wlqQLSLbjuPZBgxq3/P9lp5fGdZ2ftSVyN8YxLalqXFZpft6pHNifCOx3txir6U39U+DP5Ho+g/Sl+ZDt+23f07bI3VyT9fq9fTuxy2txKZlH9yVJVEeRa3/0YI6vF3TWA+XrlSV9lux5bpD0FdtnkPf+DuARSRO3eZ8sc+gox+P8q6QiYn3H035msudbETjIpcEqtOI9f4ntK8rz3R+YWtK3yt9eH5hL0txFaJbtJ1vca/YldtcaZxfgPidZ7kZyyPBDUr17ru3XW3iNKy3QWG+MkvQ9cvC4EVlz3EmSaS8mSTPflzTXUNC/qkf3CETSgsBMpHTgLNvnKOU4i9g+qDymdU9uxffvLuLDOAXZVEJuVhvafqzN2P2mLEZ+TBqCHGP7u+UmMR1wQpsLvkbMfpVIrkf8so52ypx2Ir5d+5AF17XAxG4hS2io0HVtb0g27fcTL+y1SOnm1MRv/i8tx/4WKVG8XtKhpLnplUQMe9X2RZI+CNDWpnWoUD5XhxB/zQvI4cKaRJD7MrHtuLB/I2wXjZ2xtAbxGd2CLP5OUDzxPmN7a/XQNkjS+uTaXsOxltiTlIi+AGD7G72I+25Csee40/a1irXYgcRG4CXiO/ksEZunc4t+mxo76/Nocvh8NakY+A2wMHmvL3MPq3MkHUtsWF4lB/JHOpmInc/A18tYtnfLJeiNGDMQu6f/2N6/zRhvYwzTl9i3lUMsJH3T9lGDOY5e0LlfFkHuF+Rw8pPAl0h1wrPAMsAfbN/RqzGQ9ebEpKnqc2RtsgTwoO0jG4+dsE0RsF+x+xFXjSzGInaeAFxsex/lAG1N4N/N2G0xFGI34hrY0vYtjce0vkZ/m3E7j1mLVMXs1VLs5jp4fbIme40I7SZ7PpN16R/biDmOMfyG9G+YgVQor0F6Ty1JbK5atz8qcZvPfVD3Xl2xx8oaVjL7XyQHH8eTfhY9mVMHA0nvIT1gIIdxz5Fq8M+Rg5Qpbf+wPHYzkmB1WNtrhPL31yWi3V0kielI4G/AD4EHbH+/PG6gmdzNNdnMpEL6FHJAdyzxnj/wzf/C+NPn2N193HYiyQ/bO31LFiBVSbZ9ZS/GUBl/ynpjP1JZ8RrxUP8emY/WJHuI2yWtDPzNLSbrjC9V6B4hdE6Au26GC5NS3VmJpcbRnce2fUImaQtS1rMGyVJakSy85wPusb11m/H6haQly6k6iqfnRGTj+muSCbejpK2I72mrC4/GQnZqYAbb/2rz779F7BlJ5+s/E2/oToOOX5EM29NIOfgtb/pHhhGNTfso4uM6NxFY/yhpI+DrpMHrgL1j3yT+e5zmsXORRd3y5PO0ECnj2tr2w72I3W8kbQPMaXtzSQcSgeQgx8f3/b16zftJWTz8jiwSfqI0gvwcMC/x2NynVwJkiT012TDtbPuGxu+WAKaw/fvOY0fCIWW/kDRlEXaXIKLUd0jG628lLUnsHTZscwPXzMIg89b85T41M2neNTkR3F9u+xCla7O+NPBt21+RNA9ZKzwH/KK8JqsQsf87tv8xwLjdwsAb83m5h85Esk9+ZPvfA4n1TsdUhPZdScnwPo3fj4jPlqSDgedsb1++X5UcWs4D7OBke7cZbwlycPCApE7j9a+VDfPypCLmReL/2eqaqV+x+/ycm2ujxYj90dPANsANtn+pJNv8twdr4KESeyaywf8YOTS8lMxlT7llG5O3GffpMoeuRPZ8W7WxFu86VFiA2GrNTMS/V0iW9bVEdOxJ5p6kqYD9bW9Uvv8+OTxbDqBtsbOshzoVA5sDn7K9bpsxxjd2uQZcfj8HsAKp6By2IjeA4j++ErkvXuZUOyHpq8Rv/5vEf/2E8vP3uOXK9PJ31yLVXhuWuXU20vDzuE7s8rgB3as1dnLLR4EHgUlIk+zLSRPh04CNScJLa3pNn2M37UoOI8lxG0jalginBzsZ9JUhRNc6fhNgF9uzlO+/RXSJbch9aMhl31eP7mFMWQDM4DSM+h8/S9tXKQ0rlu6I3OXnbYvcE5GS1I1tPwvcKOkZItB9onFzGtYbOaX0+P8aP7obGFVe969K+oekBzonoW0/33Jzmp5kS7WSrfEOYj+iZGptS97X35PsitfIqfdebYvcvTiQebs0bsZ7EmH5UmB1SbfbPlKSSZf7VkVXjTlpf1bSOiTbc2vb+yolo50mXrMSr/Bhj/43S/lm4k+N7a0kXUI2VT90GtQN63nkTfgQcK9jM4XtP0u6hTRue83Fb7MXlNfySUmX07AzKwuYYzobyRH6ug8Kivf0f5yMlZWAdcm8eRGwqlJx9UV60EyriNwiYvY0pDx1d9u7SjqTiN8ztS34juNzfRXwKUlftH2epKfIQfjUStOnTYkQPiBLi4awL5Lxd52L5U65h05g+6ESr3Xe7HNSYo9yDuz2Jh7Cbzx+uH62ujZBIpvmExsP+Q3JFJvMLVd8SZqQYgVTRImbgZsUS4ObFQ/SuW2f1Gbcfsbu53OGsdZGvyXWhNORtcgBwK5K1vgvRlLsLoHmGJLMcwzxpZ6CVPh9kTSCbM2e8R3EXVvSB4DtgbXdjh1PUxDbjqzJriSJB1uSpJcpSIVGz8rTbT8j6RVJv7S9Adl7LE0qhtchdj2t0D132z5U0tSSPm77ul6ugd5G7GYW7tfIweHPeiH4DjaO//hfiDXJzSoJZbbPJM3tTiP38s7je/WcZyNJgQ+UOPcW8XudzgPauAYan6vTyB76LlKxeiAR+wWs0K3ptEGfY3fmsl+TpsGflXSw7S0k7QpsLemHbtluqTL+dK/jbf9C0icknWB7HduHKclp73OLvRnapGZ0D2OUzrQrkMZhL9reUmOfwI9VptjWTbprYzNJuUmdRjqb/6r8/BvA7zym0/uwFkm6Fhk7A6+Tju6nklKff5CmYtMDJ7txsNBC7GZGx+bA12x/rvyup2JwI/b+5DkfQhoQnEtEmv+SCa5XfpszAh8HbiVluD29AXZd2ysQW5rDbV9chL+PEv/5nmUVl4XAkWRD8wg5+T7K9mnld+8tQs2wR2Nnm65KPIqvJM1JzibZBpuSyolzbB/St8H2EMUjeE9gXScz68PAZ23/fBDHsCHx3ryJZAU+ZHubwYo/kilC9n4kuWBPcr/YhWzabyRZeTO6Rw1WFc/DpWwvImlKIlzY9s6K1+VLPYo7iohSfyGNCf9EDm+mJo0pd3BsXKYBJnf8MduKeyZwO6mIOJJsot/we1Syxf4POKQXmzpJK5KN87nkkKMTdyLbLysH56/1UiQaTJTM7X+Q93cu8rwnJA2b13TLmVqS3tcRJIoYMTvZPH+KbNYnJoLYzrb/PBJi9/k5N9dGSwLL2t5O0tnAoSQZYBZgkrbnsX7G7hrHjqQR+RYk6eMe4hn8FJm/enIg/XbiKtV/dhKf2orbEdhfIXZT1xIh/31kL7KvW+4Z0hW/uZ89nCTZzAt8jWTeHtvWWlzSe20/rDGWlI8Tm5hVgH/ZPqeNOAONXebZbwHfsv3PXo1pMGkcqkxBKoVnI0kXs5B99eKkivirtl/u4Ti+TiolTnIqag8mdp03vMU/fScxOnvqVYAVnWaPs5AG1WeRxn6TuuWG8/2O3RjD+8geeoPy/fGkZ8emzftbZehQ5qXTgX+Rw5FdyOfyw7a/3s+xvR1qM8phSFl8QESgaUkJ33Uw5rSuI3JLeo+kbdVi44bGgnMDYGclg2QXYAtJ60v6FbBg89R1pIjchf2BBSgNK0hp1XeBHYkX5YQtxh5VbkxTEB+8q4Dzy3s6WeN0tic03rfrgT87fkvbkJPfFWw/27bIXeK+ppTWH0e8275DXu+e0Xmty9fTlcXlmeTUeW7HU/VmksXSeuzGt18gXqaPOmXHWwJbSlrLYUSI3PDG+zwBOTSZm1xX65CM18mJJ/dWxKd7WsXXeMRRNueXAyeXQ8LjSePNVpE0iZIR2PxZp1z2aCIYvEDKwLdp/r7yzulcr45dw8VEmLrL9r2ky/3qwGK2H+ulQEPuWZNK+r5TdfVTYHKlWUyrIrekBRrz2W7ks308qVr4DzkY/jup0Li23F+fbEvkLiwOXG77eyQR4BUiKndEk1WJaHROj0TuDYHNyGfpZHJY20k+eFmpxjuaiPADjdXaemMAY5iCiBOfI4eS95M14Y+BQ3sgck8GfFnSmpJ+Ql7nJ8l7fR3JNn0M2LUHgm9fYvf5OTeF5tnJ53caSdeTTMBbSJXhcz0WuQc7drPCaTRJ7HiZZFb/gqxZPkkOsloTud9p3LJ2vaNNkbswPzCtYxtyEDnE+jjx6V7WSQDp5f7j9c5rYXtTYA+S3DUzyWZvM/FlU0lfJvubi8ih9DLEJmUHpSqrV7yt2EofrA2AzYazyN28Zhoitxzbn4tJYtOswPlOctM/ScV4z0TuwmXkkHZzSecAr7QlcjfWgp099ZWkknIhp9LpGWD68jluVWjuZ+xx8BjwPsWuBHJQ+REls7uK3EOErnn9B6SKaDfSY+cw23sDjypWYUOamtE9zOha9E1FNhfzksn5GtuXSprU9n/L748DftqDhe8XyEX/NSKCrkE2N58kGQZHdI93ALE6N8LJiVXIswMa/HiOgZTb/IssrnYlJ1yX2D60POZjJBN11TYX3OUE9CSSGfc+ssicH3gP8a57sa1Y44h9IOk2fTcpD92TPP9dgD1st9rVXsmyu9D2C5I2Jgu+C0jm/NpkYd/L5zuKnGzfT06Zt1aak8xFMvUH5B37ZjEbBxYTkef8NZKZtavT2GE+4pV8Tdvx+4HiPzilY8nyaWA+28dIOp/YGRxebqRI+ggps1u5jc+V+miJ0x27ITI3qwheIRU6l7Uce2JgEeJr+lHiJ/q7zjjKgVr3+Pr2Wg13NKZaYTSZt0eRDfoepErhPpKxdbPtG3s4jgmcBsIdcfUG23t1ft5yrE+SjKxrbV9Q5vCpSYPJI0j2+nxuOQuw+zBa0kfJ4eg05DnfCmxnexNJy5JGid9s65C2e52j9HE4lzGHluuS+ftppfHWMcCPPcBmS4pP6+wk03INUlX3+ED+5tuM28y0/CbJ1H+MHEZPRrLn7yEiWat2U43P1YfIvfoe20uVA9ONS/wLerFh71fsPj/npn3GruQz9StSDfEV4Dwivu1t+7wRGLvTp+V8UrE5KXm9LyQVd9Pbvmu4x32TsUxCxNcrPKY/zWdJhc6f3K7vfGcdvAvwT9unN373xr1KqWRekYigA16PK80NLyb7qd+TnhnLldd/FBGbv0L2Jae0uSYaj9hzE//5Ng+FB5VxrYEb69//uU+0vQZ9s3tR4/qblnzO5nNpej/Q+1fjb48iB1V/IxUST5O9/AvEo3y7Hqz7+xZ7HGPpzG1TEj3lJnI4vj5ZI+3rIWp/8W5iHOvprYn13F7l+32BHd2DJJFeUDO6hxmNG8J6ZDOxCMnquA5YTPFZXUPJsDmDTBwDFrmbpztKU5KO5+ciZEN3BdnIndQQuUcN5ObQoUzSM5LS75kH+vfGk22Iz+SuZAM9C5mYl5I0Z9lwjAK+2JIYN6+kicvrvhfJyLuRCFT/Idmfj5Ibcms03+fCqSRbfWJS7r4MEQkO6YHIPQ1ZwH6j/OhGcqp/HNnMzABsq3jC94odSFbrlsBGkn5u+zjymreeTVBu/B2R+1RiDbM9KQE/DthDySa/dQSJ3MsBvwQ+KWnjch0dL+lH5DBlWWBdSe8vi7NHie1CqyK3pAFnUw40dmM+72QtnWP7ws6Ccxyfx/HGydx9lWTHb0HEv87vrGSbvi5pKiVzqPV+Du8mGiL3hSS790DgIXJYeADZ2N7apsgt6X/WdEXkHm37GdLUaZEikPbCg/FaMnctWq6haxmzRvk7ORxuvXlYea1HSdpJ0pYl1r2kcdvdxLrk9+WfjAbW6YXIXe5hEJuvM4HP216bvN8bl+thT3JAPVCRezng38DK5F4502CK3OX13pSse9YhIuQp5FDjSyQb7h5or6KvxO5UAc1Isjr/JWn1IoSdRzzCW0+G6Ffsfj5nGMtX9XhyCP9f0gzxUVLF+CrpLdCq0DyEYp9Q4i0ALEoy2yBr1VfaFpv7Fbdrj9c5gH+RJPd8Tkl6WYPczyYDPlOuyYHG7ax9OmuN84BViqDdmV87IveK5DDtl22I3IWTiE/xLMQ+D0nLl/FM6fhEH0YqOydueU30TmJPavtfbYrckmaW9F3Fc3e2tv7u/4/GveMESbuRfdb/xJa0pErVcluxu+7VCyse6NM2xiWSAHJfQ+QePUCRW43ncARJorqNHNbNQw7uRgPf74HI3c/Yo7q+74jco51kxZXJenB1srdfps34lfGnsZ4+VNKipCpxMknflvRx0gh4tjb3p72kNqMchkhailhmfI3YOnyQbHiOB+53Oo9PB/zA9nVtxGzcHFYjC68LgcOJb/KikvYj2ah/a/ybtk69JyalE5O6hSYrbzNmt13Ji+TU/Qjy2o8GPmN7pcZj/kYLSFqCTCRn2P6rpMtI1v4XiWAyOcnwON4tZjdLmtz28+XrThf7SyVtTxbd2zoNzCZyyyVkSsbGk0XsPKIIBqeQLJa/kM3NN4At24zdXPgUTiKZab8q8bZQmmPs0lbMJo3YBxLPxV2JP/JajMnmmbQXsfvId0jW9trASUVgPVTSa6Qs9gfABh7jvdhKOVtDoOlsIt8j6WLbP2vj749v7M7ij4hjnfn9Rdt/aUMg6rrGbyQZRDMC00v6UhcTawAAIABJREFUT+fz5FhdTUUWNYPa7HYkUebLq21fSg4Jf0PsOs4mBzzfIw0gZ3CsoNqK27zGNiAVKP+FNxauEzp2Yiu0FfNNWIFkcU9CDsAPARYE9iVNiy9vK1DXfXp3sqadCDjO9jcUi5JPkgP/i8rjzm9z89xYG20KLC7pSdubKTYLU0j6aXncj8vj9rD9SAuhpy9zx4HEAuVBjemZ0nq2PvzPwexu5PXelaxFNyfWapOT17h13/dG7NOBO2xvqzQ+30FpxrcgqYK6e6TE7lfcrvvG5CS55SCSAbgp8HnSQPmENuMOsdiTkb3VGURw/BlZ579Celj0qhHioMUtsd84vLL9enPdYftXysHonMBptq+UND/wwEDnmK49x37kGr+n/H/y8rBRwGsaYze1oVtoRqhUST5i+zFJM5Eq1V0p907F8/xjkrYje4IZyX1twE3gxzP2VLTbeHNGsh45l+xpH5J0UC/m7XFwEPBXkqC3J3CVpPsa99KViRf96uR6aIXG3/8msBqpFp5MSWh6nFSMv6ZUXe0IbO8BZK1q7GzqD5N9/Bm2/yDpIeKTfQu9sSnsZ2w1Yu9GqteehbHWos8BV5e5ZRngG67Z3EOJo8kc9eeyPz2GrPE+TT4Xd/dzcO+EmtE9zCgC4FREFOmUfzxNmtG87JJNbfvxtkTuEldlUX0oKWO7nAiBz0vaE5jG9lYtxhtV/j+aMd2+J1Ea0fQUSYs0TrRWk7QQEae+QhbcfyEZYr3w9tyalAFfCyyv2Do8TbLiLiULniNJg5I2Re5pgJUlzSVpC7LQXkvSok6W3kHAT5TT79Y30E7W4QykweeviMfqYiXuBeRa36zcmFuhs7Eo7/NPJH2XLKqmAG4gC7F/k9e9ZyjWNK8TG6JJbZ9Fmgy9avsY99DWYDCRtJySnb488BGyaVkLWFLSmmQzdz/ZtLcmhnVoCAadjPlVgGUUP92e8laxPSZjaVWywL+3xdidxf13gZ/Z3pOUvy8PLFGu/fnL4v4UYkk0oGzTdzk3ALspmQ8HkAXiMeTg6kRKH4c2RW4Y6xo7jVhGvLEpLnPdK+XrpZQMrlayMdTwzVdsnia2vQq5h80HyPYBwFZuMfNSYzJdR0n6DslS2t05lLxX0hmkIfaxti/qPN+2RO7m66f4FH6cNHV6RNKRtncmHtWn2/5O57EDFbk1xm/zRMXncnPbKxL/3PUlLQ5soiQItIakWRsC2JpEXL+1XFfnkHXLusSTu/V7VuP1XowcBG4L4GTm70oy1Q5x+z7FfYvdx7jNfiUzFVHicmLF8yqpmJgXWFYte8T3OXYz43PmIsK+SubvX5fY5wNPuN3M2r7E7dC4dxxb1mnqjKv8/i7bF3XWBbZvGago1bXn2Izs8z5Pmg5+i1IBU+b4L5Akn41t3zGQuA0mAg6UdDWxFvsyqeT8JFmfvUaawD9C1qXfsj1gkXsIxO4wB5mzzyF2NGeTPh6tV8zqfyvNriBr3O+RA/AngM+Ux64FbEKEz7YajTbv1TORw/hVyZ5rauAZJYHrNSXZ4yjg9wMRuWGsz9XRJIP6QLLuX5TSg0jSdOpB76E+x+5UpZwMPGz72c51Veb3zlp0OTKv7+fee4NX3hn/IFoT5XPwjO2tiQ7TegVVL6lC9zCiCK5fIRfgIaSEa0Fy6n+R0+CqJzj8m2zc9pI0h+0rSMnq/k6zktZK7cskPQsRmL9OxP1jSEn0Um3EGBcdsVXSOuQGvAwR5DYn5bgPEY/VA2xf0oMhzEvE3V+TTcznySLgVLIw2BhYr83TNMV37ylSevxXYO4ihp1GNs4/IO/BRrafaEsoKLG756BbbB9Csn1XBdYD/mj7ULfoBwhj3Yz3IWW/sxBblkmIZcrPgN843oS9bLjzABHEridWJV8gGf0ztR1LYw6QRnX/rJeUDelNwF3KQdK6xB/uO0Ts/gbpqH68e1tCNws5FHzRyWTekBzobNZmzPGJrXRC/zawRRsiaNlAfqh8PSewBHm9cTySryBC5EvlAGlpknHbqiXRYKEwZz/jA5RF4PfIxmJh2y+QOeVl0rj4224hI60RdxPF8xBSbfW8S/Zw+f2EDRFlFXL/erQhWg4kdlNs3oaI/HcWgepUIoYuK2magQoj3TQ2cnsB7yWVLz8qv9uR3D8XbDy+1YY0jdd0Q+D7xDrkMdu7kiY9F5BN3lXlcQO+h6ir0sz2fiRrfHty75qS3DNvdYuZeWWTuly51g4h98sLgDmV3iTPOQe0q7a9LtL/NtK6AXhCqXxD0qeAhWyf2Xmth3vsfj7nErdjL3As8GOlUd4dRJTbj1RPHE7mszlGQuyO2FzuI6cR0XcxUr3aEWGOI9mJvfDkHtS4ndiNb3cj94XzxzVXKkk/rey9JG3CmD3HtcC8tncie5+Dys9mIQdqkASj9d2C3VRjTdY5jJuY9ON5gKxDtyPZrofavrzMuS/bfmw4x26MYXpJnyENCGckVaxbkvfiQCLCt0a5vjuf6b2VbHaImP03Yt1yOjCq3GcWArZ2S9Xbnc9X+fr9ZD/9B2BvYGqnYfVW5F42GemF9f/Yu+9wO6pyj+O/dZIAGkgkIC0QmgUU5YgNpZhQFFC4oIJgAxFUUCAiFxQBuSBIEw4gVYoI0hEFBVQkEYL3BkEOFlpoiRRpoYVignnvH7/ZyXBIMOHM3mvP3t/P8+TJacm7Zp/ZM2vetda7fhiDKPk6j/fV9Ii4SB5MuEEepDxcXnX/pIrVnFXIGXuApeVn2JtSSodJOjml9J5G/iB5Us/+8j5nlfWHUZlFJe2UXEJ3ffn3t2p44LleIoI/NfgjPyivJD+kbicnibaQE96rlX4uNSl+Kn38BXn5+1vm9zODiLNk8fcS8g1nGzkhdJ68IeAn5SVFb2zWMcqd519L+lnpez+Vi+9Lno1XdeyhpY+PlDeN6JETzHtL2qD43huaEHvD4u/l5Qf2H0paqfE9SRs1fqZJ59ZSkrYpPr5Q0keLj3eUl9W9qYmxt5MT+usUnx9UvAbLyzXymhJ3Pm1ZVU66/0bSlsXXeir8/4cVf69QnGP7l77XlOvGgPhJrjF/iKRti6+dobmDC+s0IeaQUuyPyTMbVpSXbG5dfG8ZSe/LGPv9cgmAvSW9raLYi8h1Pd8kl1xaXH5wXr30M2tJWqT0+dAqYuf6U1ynLpL07oxtSJJ2Le4h60hqlJ36hFyzebOK4y0pz2TukTRKHhA+pXGfKH7H3yo+3rY4B97ehOPeT9Ipxcf7yrPyvibfRyt9b5WvifLA0KWNe4Q8A+XEFv6+N5XL0qwkJwy+X/reJ6s+txrHX7yux8gz/ST3Fw4rPq60b9S4LhT3jT9LmlT63hflB+cPF+2q7H5V/l0X//cJco3g9xXn8vflxOdkSVs04XebJXbOYx7QjtPkBNCGku6WNwx7h3zPGiPXYr9O0ui6xy69t5Lmzq59i9zvHyuX7VhO0rsqPs4scYuYQwZ8vovcH3u7BvQH5eewSZLWqCDuGxv/T3FNecUzR+n12ERerVLlMfeU/v+l5T7SevI1/K3F98bIg0hVv97ZYpfasKx8DW/0P3eXJ/TsLvcN3llxvHLe4Ew5gdz4/CtyX+lySZuWvr5IlW0o/b97yWWA3ign2a+Q++VHSzq9+JnRktYaZJx5va/OVNHvKuL/l3wdr/R3nTP2gLjfkSdrrScPTK4l9wlPKr6/haSJakJflD+DO2cGfO84ebBkkqSP5W7r6/3TuMmizQwYhXy3/EB1nqRh8sPd/XI5jxQRtzUrdvlr0itmMe0uJwKPrDDuzvLo4q8kPS+P9t0oPzjvIY9Av1nSXVHUdatamlvXalU5MXFcRPykOP4LJB0UFdcJL8UcJY/uv1TM0FpVThZ8Xk689kW1s7TmzA4rZgodKXe+/kteKvt7uR7T4VHxiGvjmIuPV5I7O/vJmz4eUfz9YUmfjyYsRS61Y5Q8Ay5JuiAiHk4pHSrXeK1qiWQj1sC6742vz5nRIyfINpcTlI26cVW2YSl5tun9klaRtHxEfKncjirjDYj9Sflh9Tdy4u+vEXFpSulcuVxJpbOVSnGTfC7/VU4+HiPpTrkD9vOIuLgZcRc2dvJGR5XWRyzO7+/KM8JWlu8j58tJhDsj4uAq4+VSeg/tLD8g/zAiKtkz4XW0ZUc5WXGa/GB5ulx78+9RYd3kYibtufKKiKHyQ9wmclLo3fKssc/JM3cmJJcYOzcqmhVXun6vJp9j98r3kGXlZMlmkn4bFc7wHXDPWkGuWXugpDsi4pTi6+fKSd87q4pbil/ulzVe839FxB7FLNwz5I3ivjKvfzOIuOXX+5PyTPVT5Gv5IxFxXErpLElHVvH7HRg3eUXO5vLGdLvJ1+4fp5RWlPsnF0aTakQX19Cr5WX2b5CXXf9Ovo5+SK5b/adOiJ35mOecY8Xnq8n7Y/xMftBdV76XnSE/h5wkl8Ia9JLzzLHLr/kZcr/ouog4LKW0tpy0uTQiLh1srHaIO6ANPXLS6yl5L4nN5cT66Y3+b0ppe/k9v0sMclVlMZN43+LPMnIipfHMsYE803Z9Ofm9nvwMuJ2k56rqmxav90Xys+Zf5f7om+Q+0QryAOUDVcRqs9iLyCuU14xXrvhaR77GPFJlH7x072jUiN5VnlzzTPH9URExPaX0poh4uurnj3nM5D5Nfp58svjatnKJmHdGxKED/80gY8/rfbW8pNMi4p7kMoGbyM9DPwyv+qtEjtgDX7fklR/by8/RNySXdjtH0j7hfb+2l/exuX+wsfH6pbmloXrkgdbb5OvAv8o/I2mpqGZfmSxIdLehARfoIXKCYjP5oe4iOdl7lqT9orQstqILdDn2aLnu9+ON7zV+LCosX1GKvZqcELhLnjH+aRWlUeTlTYdJ+kJUX6tsYDsaN+jV5RnzP4+IHzcpVqOzO1oeYb5CHlX+mryR2TpysmJ4FJs5VBS3cYFL8kjrafJs+U/JM4g2KP5cFxETqopbxG4c85vk0fNJxe/+NDm5+/Pi9eiJCsvxzONmXB5g2FmeQXV+NKEGYumYe+Qk59Vy4qvR6SsncDaWOybXRDXLJb8s6d5wCZb/kTdR/WjxvSPkWb7bDjbOArRjbXnmxjHJg3efk/Tn8LK6qmPtIS+Juz15IOXbEfH14uMfyzPnH5dnK1ZaTzZn7AHt2ElOOt4nP2jcKOlfcsmr6RHxP82KnUNKaVn5GvK0PGC0Z0T0tzD+d+TZxffI19J15HrcY+RZE5XXPU9e1j9efi+tL1+7G4n25eWSEjcUPzvPgbbXEbO86eXmcud4RXmWzmR5xu+zg43zWvHlBP/zcuLvJnmJ+13Rgo1lizaMlQfen5CTNrdGxJnF93atsr8w4F7dGBj7R0TsnVxP9PuSHgyXMalcKTlzW5GEGyEnqO6TV+h8J6ofFC4nSd4kz14+L6X0O3kzuAclXR5NKLOUK3bmYy73T06U+ydTJM2Qk5yXyKvdfhDes+VVyek6xy4+Plyui32JfD29ISJ+nVJaS+6LVnavzhV3Hu34mVzCbLZ879hGHrh6m5yUejGltJ08sHVHBfGGyisTNpZXsG1ZfN545lhf0oTwwOwqcpm3fw427oA2/Kj4f/dJLuW1lvzcNUPe26JppdtyxU7eeHJ7uf/3aXlF5dPJe179K5qwJ04RN0k6Vi7F+WE5ubqLfNzfkO8djzcjdqkNG8ir3XaTy8LMLL7+noi4tfRzlVxTiv9rQd5XIyX9OyouBdHq2KX+yVD5d31ORNySUlpPXo3/c7mvtmhEXFP8m8peawxOcd+9QO7PPijvCXBeeMLlnMHYKvKL2UQbTCvnz7z/yDOzzpcf4r5WfP49+aHjoCbH3lvSL+VE7wGlrzeWsY6Qk9CLVhCrMeCyuDyLeqJ8Yd5UTsLuL+lKSe9ownHOc6mt5i4ze6s8q2ZFvcYSj0G2YQn5BrGlvMz9Nrke+bDi+JdvUtweefDgYnlm2gi5o/17FSVkmhDz/XIphSXlGdyXStq8+N7ycpJqyybEXbr08ejSx41z783yDJrVmhB7SOnj/eWR7ZOK13qFAT/bKC8wquI2DJO0b/HxqfKGg43vHVL1OTbgmD8j13eXvIxul+Lj96lYstmE13yF4u8PFef57yS9t/Qaf7MZcXPHLrVhIznpu7o823druS5f74Cfq7TUQKv/lN6/w+SH5GOLzz9XvO4fbGLsgctEvyAPQL+l+PwEeQZ/5aWuBsTdXL5nrig/xF478P3ceJ2qfN3lslKnyg/qm8nLgI8pzrVFqo5Zir2nfF9cSu6YrybPQjxV0sotOOf2kPspffIM9g/Ks6v/u+rXvHR+JzmhvZe87P2aRjy5/zDokgID4pbLw2wiD3g37hWfK67dm6g5JZ/Kx3y+5pbx2lse/F9VLnHWjBI8WWLnPOZyG+R+4LHySoE9i/fyf8uJk6aVSskZu4j/8eK9/L3i8/fLE2wqLT+UO66kMaWPR8glK1aSEx2fkvvBI1VxORq9+l75efleuaSa/8wxMPY35FrFYyQNkZ+vfyBpsU6KXYq5uHy/+GTx+Tfle/fX5Fnlq1Qcr1yuZCd5tVXj8/3l/sJ1ctK5KcdcireDpLOKj/eRB8iXK66l+1QYZ0HfVys24Rhzxl5avkctIa+62bn43b6n+P4BRTuWmtf5wZ88f1R6PiiuQd8qzp3/K67Le6sJZXJz/WEzyjZVzOr8ovyw/G/5JPynfBG5NSIOKX6ukk3yitE4JRsj18v9L0mPSVq5GPVRRLxczOi5QNKUqGapfSriny4/pO8u1yBfSi5pcYKkL0fE7RXEKgcdEnM3yfhSKu0+XHx9WHip3tiIeDAqmA1Xir1eSmnT4vf3Abkm3x/lh/Vd5N95X0QcFxGPVBV3gP+WE3G7ynXbjpJHX8+TR38rU5xXo+SE1GlyR+t6OdG+XUppC3lGxxWS/lZx7B5JH08pfSuldLZ8Q5Y0Z0PKIRHxeET8IKrf7Ke8Wdvh8hLk8XLScTV5o7bRxc9uKw9m7RUR06uIXfw9PLzLdW9y+YLdJC2XUvqhJEXEQVWfY8UxD0ne3OURSaskl4S5TN5Q9o0RcXNUvMFo6Tr1cPLGsmfJ76/9JR2fUtpTfu1vnf//UsvYo1JKHyw+freceFxCfoB6WX64ukOl93UxSl/bWRXFeyuK68rp8tLfRSUpIn4mJ0G/nIrd3iuOPaT0vj4ieSnmQ/KD1JHJ5SWWk/ST4r1XVdydk8sPKaV0VErps/IM8m/LtT4ny/eQFcv/Looe7SBjl/uLb5H0z4j4mlzDbwf59b5I3ghxZhUx5xFX8us8S579+Q15IGd9eQb/oDdxXQArybOkHpQTQpPlVTqvWGk22OMvzrHGTNfGhr3PhVf5fFLSpimlAyLiuaiwTEtpZnFKKa0e3rj22pTScfKqkI/IK7GujWKGbYWxFyu9bjvL/aDGZot/lq9fv5R0alRYoiVn7MzHXH5++Ki8CuU0eYLLM/Lkg0nyZISrOiX2gHZsLD9r/EzSzOJa3i9fyyrti+aMm1x+aIeU0jdTSsdKeo+ccJ0on1+T5bKR74iIhyqMW75Xfi15pexv5Jn7RxV//0S+p1RqQOzditjnyc+U35UHkC6Qy+C81CmxS21YVE50PyPPoldEHCdPtpmhJpRKaVzLiv7o7yXdmlI6uPje4fI9e4eIuKKq/MW8pJSGy5NNliliHyOXc9tX0rTi8yriLMz7qtLVwjljS1LRF3mDXGbq0xFxlnx+n1pcz94l71vyZOnf1HdmcAdI3rh+q5TS55JXof5KfmY4Wh50/aVcevDN+VpZLUqXtImBSwOKZMUX5YvyEHn264QoLTGqajlBcUPYRD7hV5Vn1R4v6SX5ZjheHv39vZxsP0euBTmo3d7LD1TFA93Okh6OiGtSSmvInZ/jogmlDUptaCzL/V2UlhsXSe5ZxcebyEsL/1Dhw/ui8ih7r3zTvya5ft3X5Nm+n5R0ckRMqyJeEfMVS9eLJMn68oYvkbys7t9yfdNK6zGllDaUNCIifpVS+rY8s3WriHgouUbcODkBfWDFD+2Nc2ukvIHXsIj44Hx+dlv5xnxwlQnA4hw7SR5UmCEnSM6Rl9N9SZ4NuETxtR2rOP7ScY+WBxMOiYj7Ukony53e/eUOySGSnqjwvP6sPHPgqJTSZfID25CIOCCltI28QuJQeWbaA1XELMUuL/E/WH6YelheHbGf5s7+nBXVLzvPFruIv5h8jo2W69t/Q555uoy8Od9DxYBHZXsbDLye5JBSWlLuoE2IiJ8Vnf2hckmJdSV9o+rBq1LsxizEZ+WBhD3kjmKPvGTzxxFxXYXxhkj6ulyGpkc+xklF3G/JiaHvyuWBqq71Xi5X8n5JT8orYL4dEU+klHaTZ4AcW3Hc90fEn4q4X9fcvUmukJNCR8urRI6LiF9XGbuIX66NvWREPJVS6pPfZ4+FSxKdJJeUuLb4ucqWeRbHfbo8aPainAQ8JiJ+kTyIuGZUWIt+wO/5Qnn5c49cMm5mSumd8v3qwIj4Q1Vxi9iryA/qk+X+z0z5get6eQKE5Nd9ZFRfnzpL7MzHXD63F5OfM7aTVxWeIU9y2VbuGz/QKbHn0ZYPyhvj7VbE31JOhP20U+KW+oNLyfeqZyNi3eJ7ffIg3Vj5OnpNE+L3yJMcHpavY3fJk102kFckfTkq3L9iAWOvJ7/mX4mIFzsw9qJyGbNHJN0s9xNmFMnmZsQrl+NZV/69TpJ0u9wfWiwi9m9G7Hm0pVcekL5FnlX8YOO4kwcWXyo+HlQJjZzvq8yxy9fvD8h9s+GSPlv0E7aWz/Grq+wDoxrF++NiuQTexsXXflB8+53y4MRNudpXuWiDaeX8mftHLtexWvFnP3npyfLyQ8fXmhQzyZuB3CzPFpdcz+vvcgL8GPliKbljMujlqppbGmQ5OeG2jryE4lh52f3n5OXBzdjRvbwsd0N5RsFKckL/RLkeduP7n1ax1KwJsT8tz5DaqfS1n8odozUrPuYhjfhyMuTj8gyKXeVE6IbyzJIDJX2iSbGHq1jaL88u+LE8+vxBuX7b0KrP6+LvofISqxXl5d/fKL6+soolPMX77LeqaGlwI3bx8TbyDK1F5IfWb8hL0N+k0vIglcqrVNSGN8qb+3y+eA1WLb5+ioqlslX/kTdUOl4exPl2ccwHyyPGjfNgpWbEbrzucsL3pOJ83qD4Pf9JxU7znRR7wHm2pjwD8LDS1/aTk2OVLgvW3Ot3j7yD/YaSlmnm6zswdvHxEvIA7ZGlr20qJ0maVtqg+Pjj8gPs6OLz9eVBMql0D6kyrrwiZnv54bERd4NS3OWa+LonuRzMt4vPv1mc53vJA1rrVxxvtLyi61Nyv+Do4hq9gzyA8z/yjMDNm3S8a0n6VPHxlzR348+l5Qf4zxXtOr7iuOXz+2OS/qGiJIucFL1Znj3VlN9zEedYFSWW5EHRn8nLn/dp5nW0uFc8IddjltxPOFjSVlVfw9olduZj7pETy6fL/e7VivfbD+WZ5E0p15crtl55/f6SPLFkcXkV5T7F1z+ioq9U97jF/ztkwN/vkcvj7F18vqT8/LNUxXHLx7yuvBGh5OTjhfLK3TFqwj1rIWIv20mxS3EXLf7eRr5PbyaveDtOA0psNSH2h4u/NytibyI/e/1ARVm3Zv+R8xnflleKjJYn1xwzv9/T64yR5X3VBrHLff/tVJRNk59rfy4nuDeT9MZW/K75s3DnTPHxKHlyzMnyREvJ/d3DJH00d1ur/jNUyGrAKOjH5aXtt8mzjB+QH6qOl0ckT604dk9EzI6ISCndJXe2nylm/l2YUvqnPHvriYg4QpKi2NhqsGLuJoB9ku6V65v+Wn6AX0Ue7f52VLiETnrVjKUl5IfIM+Sb4llyTdsvSzqhmIHa2G180DOri9/17JTSCvLv+Sg5KXNhMat+opz8vS0q3swr5s42PUpeavSU3Bk4TZ619ZWiTRvLgw6/qiJu8oY6WyRvkLGKfG6fIK9WeIOciFtF0m5R4YyOAb/n8yWFvMHRqZL2SS7h8T5Ju6WUPiI/wO8UFS0NLr2ne+VZiFfI59jx8sDK1vIM92mlf1PFxpNzRtoj4oWU0m/lJPcFku5NKY2MiN1SSssNNta8RMQDKaVj5PNsmfDGLwcXo8WNJZQPFW1txgYXn5MHTL5azCr5uvyab6fmL8VqeezSefYVeTXGx+UyKd8ML1GdLGlqRDxVcdzye+thOQm2TUrpiGjiZsGl9/UK8rLUB+SHufMaxxwRv2tS7PKGsUkeRDpU0lHJZWlmSXpX8nLSSmdpld4nG8uDo6cVcfeQN5h6dxG3mTujbyxvatvoCxyXvCljkgfhJ1UVKKU0NLwK4SA5kR7yPWq4nHB+Y0R8r6p489Ej6dPFrNteSUfISbiX5CRkr3yvPqdo86CvZwPuW8PlMkcHS/pBSmn/iLi1eK9XWo5nHqszFpePTxFxeErpEPmc+0JEzGrStVtFjKUlvZBSelNE/L5YtbG2XNqtmXLFbmncAbMYfyoPxD4kJ8EOlwethkruN3ZobMnPHaPl68k18kCxovqVClniFrFT0ffvkXRRSulm+R6xY/H5cvJg/H5RKi9QhVLfZDNJkyPi/1JKp8l9wxnyffvKqHizyYWMXXlfJWfsIu4wSXumlKZGxMXJZSHXl+/TR8mrk6uMV57d+25J26aU3hwRvyyuY5+UB6wPjopXms2jLevKEwAuSym9KOm9cr/4x3IZwTkGc//K+b7KGVua0/fvkfv+z0paK6X0SET8KKX0sjygcWBEvFB1bLw+A86ZH8irSi6VtJikbyVXUJCk/wnPyK/35pMDULpEUkppL0k/qyLRtJBxy0nuj8nL+n8v6Tl51P+v8jLdGVEsOW/GCZhcOuIOeZTng/KSo+/Jo7C3lB7uB71TbulBbqg8E3BT+cb/Djn5d5dKle1GAAAgAElEQVScZH25ysRnEXvODrLym7xRX+r0iJiUUnq7PHNpj4j435TSdyVdHBXWEU4pLS4n9I+Tk5/LyKViTio+3iMqXKo54BzbUdLREbFM8flO8kzMfYo2rCMvvd8jKqyHnlI6QU5S7BkRp6eUviTPuNxF0uOSFq86EVfETXJpjufCpTSOlvSCnDhYT35f9ScvvXo0KqjvOuD1/oC8EeMV8sDC++WZHMfI53elybA0t3zGcvKsx1/J15Qn5WT72+VBnF2jwrrBA9rwGTmZ/Xe5BvwT8gqF8+R6gPdXHG9gOZ415ITUfhFxV0ppf7kczckRcUOV18/Msd8m6aGIeD6l9Dk5mb5DMbgxSp61P0zSzTF3yWZlibji4w/JM0l+Im8WPE1OSO7fjIfX0r3jzfI19Gx5JuBV8nX7F5J+HREnVR273AZ5EPoR+T39K3mp37fk4/9BxQnf8vVkZXmAcKqcZN9cLm12t6QTovrlqQPP79Xk82rviPh78iDeiIi4vuK45T7C++RryP6Sro2I85OX6v5UHkiaNtg+yfziFx+/XX6N/xkR300pLSOXSrkuPJDU+DdVvrd75HN6ktwvOk4egP8vSd+psn9QxGv0i3rk8+uZiLg8ecCyJyL2Ln7uLRFxTzNiFx/vLG8GfUBK6XtywnOSPNPo8qh+/4wssdvhmIu+0dryoM1U+Vp6p3y+XSjptxExo1Nil9rQI1877peTkL8qnn/eKScAToyIIzslbhE7yf2SZ+SE44Xy88YZ8jPXg1HhMvV53LN2lAekG6tz3iX//r8XpVKcxK6kDYsUiaqN5STvlOJafqpcPu/YKhOQA97T75X0vDyp44PyM8Cj8mSmAyLi7qriDoxf+nycXK7j1nB5r2/Kzz0nRsTf5/VvBhNbLXxf5Y6dUhoVxd5RKaWt5GfLU+R94+6WBymPlbREM57n8fqklMZEMaEupXS1pJvkvYxCngD3rLzK/9KI+E22hjZR1ye6k+sc7ippVfkG9EyGNuwgP7TdJyd6j5dH5L4q6YKI+N/i56q6QI+RyyT8OaW0q3xD/n0Re4K8hG4/Sb+PiH0GG28e8UfKJUuGy0tzn5DfcKvJM7lPiIjnKo5Zfng9RdL04sF1I3mH2YPkpMkdUdT6TJ5VNuhk+8D/J3mW2rNyx2CYPMv5CDlpUNn5NzBRUXzth3Ky9TNFMmE3STdExN+SZ+dNrSIZWT7m5JkN28vXmx2Lr31dXtK2bRWvcSlu+fd8pDxg9P2IOKH42iVyaYf+0r+p6n1VnvG5YURcn1L6qJzgvkEeTFhf0m8i4t7BxptPG1aUH6j+Tx6tvU1ONL9NXqa0a0T8tcJ4AzuaO8illvrlGfRnyiVa9ouICVXFLccuHiJPlDfle1Se3fBZ+TXfRk4azQxvlNIJsVeRB4kOLxLbu8gPUTfL76kp8mDKalHtgFW5Dvmuks6PiBnJs/X/JHd0vyMPplW6cXCpDUvKq2yulTRdng35snzfOlbSm6NJq4CKjw+Q71tHyjOcPyKXQ9pS3mfgq1FRHfQBcbeXE/mryr/jZ+Q68JvIs6wHtV/GPGLP6/x+WL5X7yK/3uMl7R4RE5sQN8nHe2dE7JdSepekPeV9Ms6r6t78H9qym9wnekQetPt1eKbY0pI+HxF9FcYq/673lGsW/0jul31Xfm9/StJfIuIvTYp7jbyibpikRcIrU06T9K+I2LOq++R8Yn9Avk+8KyJ+WHxtFzkReEUT7h1ZYrfRMV8iaVl5EP5RORl0tjyo8r2IqHTD5MyxywnIX8rll56WN/48R55dO714FjxInnk66M0Bc8Ut4pX7ogfJia+zUkq/kxNjW8sDh0dVEa8Ut/x73k7+Ha8iP+dNl6+lK8gDh78ldqVtWEFzE4+Lys8Am8qJ52XliUb/qDBe+V59sTxz+na5ZOIV8mSbdeVSKROrijuftnxW7geGfE3ZQB5Ee5+k+yPi+IriZHlftUHsxeTnmkXl3NSR8krCozV3cuZn5GftB6qOj9cneW+0D8kl/x6Q8z6nppQmyu+Tu+SVRfc03stV9vHaRdeXLikSBZfKMz8PTCkd1uzRqAEdoDXlB7gvyqOh28oXjIskHRERD5faWkUybjH54XhUEXsbeVbvT+RR2GGSLpFnAlYye6d4WN6ndAF+v/ywfKj8BvywnJA8U06SzKwibil+OUGzhXwjfkvyMtHrkpc7zZAT7DOLn1NVD9IR8XLRCdldTgb9QV7mfqb8WnxBfrCsMsm9SriMRI9cN2uKpCTPjPuafGP8r4g4pdTOiVXFL455tHxu/T0idkopHZBSOk8+t/8kr6KoNFkRc5d9ryh3vj4i6bMppVnyjtsryhs+lf9NFe+rntI59kVJyxdJuSvlZeZ7yYmLc6PiQZxyG+RlsCfLZXD+KNdqW0yeSfTJ8vWkgnjljtcukh6JiAuSR/s3kM/xXeQ65JWUhBkYuzjmY+Tz+ya5/NCWcrJ/U3kQ6wPyLtPnRAXLoTPHXkwekHtW0rjkTV+Olh+mloiIHVNKf5R0VUT8ufg3VczkLl9PDpLLGrw9pXSgfG7tJK+O+W7VSe7SA9UQeRBlY3mFwD5ywn0F+bp2arOS3MX7egV5VtRoedbrZSml98oPWGfJ77MfpJS+FYNcMTHgnrWZPFC1i1zPdqg8ED1EHgivJDkyj9jzOr83l2cOrS1vpFXl7PXywOxWkp6MiP0kKSL+mrxZ8t4ppd9HxCNVxS3FL/fLNpMf1HcIl1A5TtI3U0ojIuJsueRaVe+t8jm2olzK4cPyffsQ+YH985LOiwpXAQ04x94j6c9RbBaWUjoupXRgkex+l1TNfbIUu3zMu8nn0wqS1kgpPS8nTa6Q9NMm9AezxM58zKkU++OSboyIvpTS3vI597LcJ9ynSYnmnLEbybhl5dWbv04pXSyvlv2ApJ7k0npvkRNjg34uzhW3ofS+Hi2Xuto9pfSM5pZHXEEejK/MgOvJVpLWKOKcKidfvyU/+/yk6qRKt8YutWG4fK+4Sj6/tpQTXKfKfbMfVpzkLq/u/qy8EeKXk1c87SZpWETskVJ6WzR5JnfxvLGrnLAbLd/LLpInCs5oJLmreN1zvK9yxy5+1y+llG6UB+umNHJkyeVtx8mTFA8hyd12/irfXzaVdEmR5D5AzvFNkDdqndO368Qkt+QahF0ppbR9SmlkSulNcsLxTjkxc0iRoGpW3PIFej15Seof5RmvT8uztFaS9N5GUqq4uFUV+yV5mcsb5VIlh8pLYmfK9Sc3LmLfU2Hsg+TZy0opfSAirpVvSnvKnfv/lTu7w6ru4EtzbhA9cgJwN7nT8RdJZ6SUNpVnSo1sxI7CYOMW59hixc3/UPn8GiWPvP5FHhk9Xk7uV1a7LKX0frkO9Try7JG/y/Wh/yHptOLGf2fy0vNKpZR2TSmtVXS8zpaX4H4xpbRjRHxf3izv63IS4+kK45bP0x3li/iz8nl2g6RPyMnmL0XE7cX5UJnSg9yZclJilpyU21KeCTlcTgRXvVLhKymlNZJL4vxM7uwMl2e3bi93iFaXl8tWluSW5r6vUkqHyXXeP5hS2jwirpBrjm4mdzQrTXKXYic54Sf5vNpdPs+fk2dxfFeuw/85SYdWkWjOGXvA9XsxucP7mDxwdXVEnJNSOltSfyPJXbR3sIm48vVkvNxvGC/PUjpU7jRdIs8wr3ypZvFwuLg8w/cleSDhQHlQ5+vyg+QuVV5PSrEb7+ufyffH1eRk7y7Jtf03kzeVCnl2+YGDTXIXcRvn2Dnyg+pM+b7xVbl+8hvk93SlSe4Bsed1fr8o6eGIOKnKJHcpbk/ySpxhkmallNaXpJTSByUtJe/l0Owk97vkRNSVUQychFfB9MmDC+U2D7qfMOAcO1pOgO4kX0sekQdsp1SZ5C7iNn7Pl8kJyDVTSocW375I7iM2jr1SMbfO58/kVQpT5EHhyfK9c7S8AXgz+oNZYmc+5kbS9Wx5g7j3JE90OU3ug/9bnu1V+WzTNol9vvz+XT+ldJZcJ/sMebbrzcVrPkXS56KCsim54s6jHzxRXil8i5xM/5Sk6yVNjGKlcFVKz1mnyBvTPSP3R3eX71kjJd3UjKRKt8aWpORSblvKE5nuC6/4+bm8gvahiDgsKiy/Kc29liWv5psp7y+wbkQ8JvcRVih+rtlJ7rXkMpQXh1dNnl60556I2D0i9h34b15vzNKnLX1f5YwtvaJ/8n15lcBfUkpfK759oVyyry8iJlcdG4NTDEisX/z5fHLJy8lyXuY8OR9U6bWhHXVlojultKi8FPR5ORn0QEQcLdfunS7PyhrRjNilC/Re8oPktnKSaJZ8Y5wqn3x/GPhvqootJ7bfJ8/g/rA8q/hWeWTurihtOFlR7LMlDU0p/Vye3bhBuDzIVfKNaYaks6O5ZWO+J98Ad5GX39wq//4PlJdL/il5xmAlSufYy3LS8d0R8T8RcbLc2T1ZTnp/LKpf5n+HfOPbUp5R/UxE/LtIcD+SUnpDROwXpfIdVUiebTpLrhf8Q7kcy9nyztfrpZTGR8Sxkj4VFZfuKL2vVpdf3+vlY19Orv34XXlm9YeKn6+0rmvhUDkR91H5OvJvefbtKZKOjIg7qwxWvN7/kpMiF8jHPE2unfaynBTbVtIpTXxvnSFfvy6Q9EtJ2yUv+V9Z0s+jqOnWJAfKM10vkQc0+uVO4Nmau/njE3I5ib/XPfY8rt8byOf6LElHp5TeIZfF2V2qboBUc68nn5ATgNPDMzdOkx/ixku6LCImVBhzjuK6/Eb5PbWvXHM+FfEfk7RvVDyIM8D3Jb0QEdvLScd75PJPuxWx75C8iqXi99n3Jc2KiM/ICbG/yw+Qx8p135tSHqbwWuf3qCbGPUO+dv1Cvn/uklL6ljyLfLFo0gZHpfvHnnIf4VpJw4sEu5JXIg2LiDOaEV+vPMcukhMVN8t7KhwQzVv63Rh0PlR+YH538l4WZ8mDpM20u3xtOVy+ttwgTzi5OSIOjog/dWDsnMd8iHyObSxP9thELq9wgrzS7I4OjV1+b50r95FmyZOKftwYyImIF6PaDQJbHnc+/eCN5IHRd8oD7zuEJxs1w0Fyn7Qx2eM5eQDncEnHN6Ef1tWxU0rLyhuaPioPir8npfQ+ebB0o5TS8Gb0yQqNvv8v5RIIX0kpfUNeJV75YHTDgBzKLvI+KUunlNYp3kcflMshqfi5KmZyZ3tftcF7WnJ/oD8itpOv2VuklL4t9x9OjIjfNfE8w+uUUjpYzq/tIj8/bSHnfU+T9I1owgbI7agrE90R8a/wjOVxcm3Td6eUtg3Pqn1Evngv1qz4yXV0PxQRW8kn38fk5d/PShoXxYylqi8cyRaTk1/flzvd/5KTc+tKemNEHNaE2A/KsxuXkWsSrl3Mhkvy6/1MNGFm2gDXy4MIT8qzLDaRExZXSvp0SmnpqHCX99I5toE8k3pkcn1sRcSP5M7+3c1I0IRnhlwl12d+UdKGKaWvFjMzP6FitL0JcV9SUfZGTkDNlmcB7ig/0K2aXC6mkhq2AyVvALOz3MlbXC7b8SM5GfekPKN7i+RVHFXHXlT+PS8jLxW6TL6+3inpy1FxvU1pzut9qZyEGinpf4s4V8lJqWsl7RUVbwDZUMxiuUdekXGWpHfLSd73SzouIm5pRtwi9mKS/inP8lxErg39VrkE08kRcXPykrt/RMSDnRB7HtfvPeTO7lR5UOW9EXFh8bM9FQ6QNq4nf5HvUZskb2y7pjzzdqjmDiBVPVtpqFwf+UPycT4rv8/eI68+OqaZMxJK7+vlUkrviYi/yQ8bl0jaqVkPF6W4yxZx/y4PnJwrabOI+H0z4hax/+P53aS4jevJjfJD9Bvk6/hseSb3Vc2IW4o/Rh5EOlrup7xJnvRwjqTnm3U9m8c59vfi8wnyQEozH2BvlfRgSmmZYvD3B/Ksrc3DG+Y18+H1/+RJB5fIg+IvyEvQF2lizNyxs8QtzrEHJS2TUnqr3G+YJc/kHxJNXHLeBrEb19HeYnBwtjyQtmFE/K6T4hax59cPXlIerHugWbHl56wH5Y2afyqX8bhbLsM5sYlxuzV2r6Sr5UGUbeVSjYfIq7/2iojnq+6TSa+6V58mr3QaIfcZvtXs17u4V2+luffq5SUdkVwK6N6ocEVjKWa291Xm2I3f9Z+TSy+Nk/tmb5P3LXlJ6tyyFzV3vXwfalyX1pOrK9wZEX/M2bBW6urNKFNKS8izL3eWE0W3yQ/vO0aF9azmE3txeQn0lvImLIfJM3f+/Jr/sJrYO8kdzQvkMimnyLM+ryy+34waYkvLScC95ZnVf5L0FUlfbGaiohR/GXmGayNZsqt8AThHrhM4KZqzJHqknKR5r5xcvzMidqs6znxiD5cf3LeQj/EWSb+LiGuaHHdxeUn/u+TR9b/JHb+vR8TUJsceKZfq+G95xtAe5cRMauImZkWCaGv5WC9WMcBRdaJ1PnG3l5cj/UR+3beXZ843e8O2VeXZrU/Jg2bflTe9qbyMxDxiLyaf32+XBxbukTSqGe/jNou9k+Zev1eWO7xHNXt0vriebC1fT0bLD3InyQO1z0fESU2Ku5r8nvqE3NE/Rp5FtFKzrydF/Mb7egt5wG4neXPVSkt3tEvcUuyWn98Dricz5Zlyu7foerK4fP16Qu6rXC/ft69tPBSkV9YmrTJ2rnNsabn00ePy672rpO+0YqZPSmmUXCNySXkAZ195gPQ3nRo78zEP7J9MlbRks/snbRS7m67fr9kPbmLcxnPWNPka/lW5TvVQSddExG3ErjTuaHmPp2XkvtgWKiafNPueOY979YHyM16r79Vvlt9f75WP+3fFzzQjj5HlfdUGscu/65fke9ge0dxV+BikAdel6cXHB0TFq/nbXVcnuiUppfQG+eYwVt6V9PxWJCuK2J+QEwbvlPRURHyvRXFXkjtds+Xk2N2N2M24OQyI/WG5puqx8uZHldad/A+xV5RnWG8tb2S2uVzW4r+jgrqqrxF3iSLuenLJgS+Ea5k1XfKO7h+XZ9ue1ooHiyJuYyDns/KN8aBWDGiU4q8ib1A3Z9lgY3Zak8/v4fLSvXfJm3M0Zfb6fOJ+Uq4R+Kg8E7DSUin/If6a8pK2UyLi5y2Mm+X1zhl7HtfvKRFxUPG9Zl+/y9eTk+WB0qPkGbfNXH7e6GyfINcn3KeZ1+x5xH6DPHi0pVy7+exmv9Y54xaxc7631pT3rzglIi5vYdxR8izTx1NKP5X0l4g4pvhes99buc6xFeQVfePkclOVrz56jdjLyn2iDeWyS9d1euzMx9x198sidtddv4v4q2hAP7gFMVeUz+2tJZ0or/D8lDz78yliVx63sfHpxnKf7KOtmABQit8u9+rbIuKHxfeafa9eRS1+X7VJ7CzPeXj9BlyXTo/mrhBsS12f6JbmJCI3lDd3eqKFcVeUT75VImKf4muteohdSp5dsU5EnNji2OvLm2Y0s67qa8UfItcmP17S9tGEDTPmEXOE/DB5S0Q82ex4A2IvLml4VFt/cEHiDpffVze1+piL+ItGhZt8LkTc4ZJGtGrArBT3DfLgwuQoNlFrYezlJC0XEf2tfJArYmd5vXPGznz9XkIuc/VoSmlrSX+Nimvuv0bsFST9u9XXsiL2G+VB8XUlndeqWRG54haxc53fOa8nPXJZt9UiYu9WxS1i5/xdN2W2OrHbLm7X3S+L2F13/S7i5+oHLxLFxqqpiSspiT2nT7avfH5Vvvn7f4jdrffqLO+rnLFz/q4xOCmlYa2cHNROSHQXcna0S23IduHototWco2vRcJ1tFsVM/s51mrdeMw5ddv7GMbvvTWKQcNNJN0YEY93etxulVJasjELr9X3sIznWFf2P3PF5prdely/W6cb39M5Y7c6od8uct6rgbro5v4GiW4AAIAFQGKse/C7BjoL72mg8/D+AjAvJLoBAAAAAAAAALXWk7sBAAAAAAAAAAAMRlskulNKZ6WUHksp/S13WwAAAAAAAAAA9dIWiW5JP5G0We5GAAAAAAAAAADqpy0S3RFxvaTpudsBAAAAAAAAAKiftkh0AwAAAAAAAADweg3N3YDXa+zYsdHqmH19fZKk8ePHtzp0V8buxmPOGbsbjzln7G485pyxu/GYc8buxmPOGbsbjzln7G485pyxu/GYc8buxmPu1tjdeMw5Y3fjMeeM3Y3HnDu2JE2cODFlCdxcLc89ls2aNUsnnnii7rzzznl+PyJ0zz33tKw9EyZMGOx/0fRzpLaJbgAAAAAAAADoRJMnT9aVV16Zuxm1QqIbAAAAtdTb26uJEyfmbgYAAABQuVtuuSV3E2qnLWp0p5QukPS/kt6eUnowpfTl3G0CAAAAAAAAgBw22GCD3E2onbaY0R0RO+RuAwAAAAAAAAC0g5EjR+ZuQu20RaIbAAAAWFj9/f3ZNrgCAAAAmmn11VfXtddeK0lKKSni1XtjbrLJJq1uVlsj0Q0AAIBByVknmxrdAAAA6ESzZ8/WhRdeqPvuu0+SFBGvSHbPK/Hd7Uh0AwAAYFDGjh3b8ph9fX3q7e1teVwAAACgFX7xi1/ojDPOyN2MWiHRDQAAgEFhVjUAAABQrcZMbiw4Et0AAAAYFGZ0AwAAANXaYost9Otf/zp3M2qFRDcWSG9vb9fN1sp5zP39/VniAgAAAAAAIL/HH388dxNqh0Q3Fkh/f7/Gjx/f0ph9fX0tjTdQjmOW8h83AAAAAAAA8vrIRz6iM888U0899ZRSSurp6ZnzvZSSUkraa6+9Mraw/ZDoBgAAQC1140A8AAAAusNTTz2lXXfdVbNnz87dlNog0Q0AAIBa6sbSagAAAOgOL7zwAknuhUSiGwAAALXEjG4AAAB0qje/+c1aZZVV9MADD+RuSm2Q6AYAAEAtMaMbAAAAnerYY48lyb2QSHQDAACglpjRDQAAgE618cYb6ze/+U3uZtRKz3/+EQAAAAAAAABAq4wYMSJ3E2qHRDcAAAAAAAAAtJFlllkmdxNqh9IlAAAAAAAAANBCl1xyiU4++eTczegoJLoBAABQS2xGCQAAgLpaddVVczeh45DoBgAAQC2xGSUAAADq6q9//WvuJnQcEt0AAAAYlJyzqpnRDQAAgDr60Ic+pJ/+9Ke5m9FRSHQDAABgUMaOHdvymH19fert7W15XAAAAKAKa6yxhiZMmDDf7//73//WmWeeqSlTpiilpJSSJM35W5ImT57c9HbWCYluAAAAAAAAAGgjd999ty644ILczagVEt0AAACoJWp0AwAAoFMNGzYsdxNqh0Q3AAAAAAAAALTQ1VdfraOOOip3MzoKiW4AAAAMSq4NIfv7+7PEBQAAAAbrj3/8Y+4mdBwS3QAAABiUnJtR5kqyAwAAAIOx4YYbatKkSbmb0VFIdAMAAKCWqNENAACAutp000216aabzvf7999/v3beeecWtqj+enI3AAAAAAAAAAAwF2X6Fh4zurFAunFpcM5j5mIGAMB/1o39EwAAAHSHVVZZJXcTaodENxZINy4NznHMUv7jBgCgLrqxfwIAAIDOEBH6xz/+IUlKKb3q75deeilb2+qKRDcWSDfOmGJGNwBgYXXj/TKnnK83v2cAAAAMxsknn6xLL700dzM6ColuLJBunDHFjG4AwMLqxvtltxo7dmzLY/K7BgAA6BxbbbWVJk+erIiQpFf9/fDDD2drW12R6AYAAEAtMbAAAACAurr99tvnlC5BNUh0Y4F041JsSpcAANDeurF/AgAAgM4wbNiw3E3oOCS6sUC6ccYUpUsAAAAAAADQDBtttJE22mij+X5/woQJOuSQQ1rYovoj0Q0AAIBa6saBeAAAAHSH973vfbmbUDskugEAADAoOcuHULoEAAAAnWjatGm5m1A7JLoBAAAwKGPHjm15zMbMamZ0AwAAoBOtscYaWnXVVXX//ffnbkptkOjGAunGzZ7YjBIAgPbWjf0TAAAAdIe7776bJPdCItGNBdKNNTDZjBIAgPbWjf0TAAAAdIfhw4fnbkLtkOjGAunGGVPM6AYAoL11Y/8EAAAA3WHGjBm5m1A7JLqxQLpxxhQzugEAAAAAAJDDyJEjczehdkh0Y4F044wpZnQDAAAAAAAghxtvvDF3E2qHRDcWCDO6Wyf3cQMAAAAAAKC5HnzwQR1wwAGKCKWUXvG9lJIeeOCBPA2rMRLdAAAAqKVuHIgHAABAZ/jFL36hqVOn5m5GR+nJ3QAAAAAAAAAA6CZLLrlk7iZ0HBLdAAAAAAAAANBCa621Vu4mdBxKlwAAAKCWunGzbAAAAHSGtddeWxMmTJjv9/v7+/XNb36zhS2qPxLdAAAAqCVqdAMAAKBTLbfccrmbUDskugEAAFBLzOgGAABAp3rppZdyN6F2SHQDAACglpjRDQAAgE619NJLa+WVV9bUqVNzN6U2SHQDAACglpjRDQAAgE41dOhQjRkzRs8999yrvhcRkqSnnnqq1c1qayS6AQAAAAAAAKCNXH311brhhhtyN6NWenI3AAAAAAAAAAAw19vf/vbcTagdEt0AAAAAAAAA0EZWWGEFDRs2LHczaoVENwAAAAAAAAC0kUmTJmnWrFm5m1Er1OgGAABALfX392v8+PEtjdnX19fSeAAAAOhOU6ZMyd2E2iHRDQAAgFrq7e3VxIkTczcDAAAAWGgRoRkzZigi5vn9z3zmM7riiita3Kp6I9ENAF0oZ3Kov78/S1wAnYcZ3QAAAKirH//4x7rgggtyN6OjkOgGgC6UIzkkkSACUC1mdAMAAKCuNtlkE/3qV7+a74zuGTNmtLhF9UeiGwAAAAAAAABa6PHHH9dzzz2XuxkdhUQ3Fkg3zpiitAMAAO0tZ+mSbusXAQAAoFoPP/xw7iZ0HBLdWCDdWAOT0g4AAGB+xo4d2/KY9BEAAAA6xzbbbKNtttlmvt8/66yzdO65578oga8AACAASURBVLawRfVHohsAAKAi3bgCKidebwAAAHSqDTfckET3QiLRDQAAUJFuXAGVE683AAAAOtVLL72Uuwm1Q6IbAAAAtcSMbgAAAHSqd77znTrooIP00EMPzfdnzjzzzBa2qP2R6AYAAAAAAACANvPMM8/o0UcfnfN5SkkppYwtam8kurFAunHGVM5j7u/vzxIXAAAAAAAA+Z188sm69NJLczejVkh0Y4F0Yw3MHMcs5T9uAADqohv7JwAAAOgOn/70p0l0LyQS3QAAAKilblxxBgAAgO6w7LLLasKECa/5M+PGjWtRa+qBRDcAAAAAAAAAtJGXX35ZZ5xxhu6+++5XfJ0a3fNHohsAAAAAAAAA2sjVV1+tiy66KHczaoVENwAAAGqJGt0AAADoVEsuuWTuJtQOiW4AAICKUDO6tXi9AQAA0Kn6+/tzN6F2SHQDAABUhBnGAAAAAKqwwQYb6LLLLsvdjFoh0Q0AAIBaYmABAAAAnWrWrFm5m1A7PbkbAAAAAAAAAACYixrdC48Z3QAAAKglanQDAACgU73wwgu5m1A7zOgGAAAAAAAAgDby4osv5m5C7ZDoBgAAAAAAAIA2MmLEiNxNqB1KlwAAAAAAAABAC7344ou67LLL5vv9mTNntrA1nYFENwAAAGqpv79f48ePb2nMvr6+lsYDAABAZ/rRj36kq666KnczOgqJbgAAAAAAAABood13311LLrnkq74eEZKkqVOn6sYbb2x1s2qNRDcAAAAAAAAAtNDw4cO1yy67zPf706ZNI9G9kEh0AwAAoJZ6e3s1ceLE3M0AAAAAKjdkyJDcTagdEt0AAAAAAAAA0EZGjx6tCRMmvObPjBs3rkWtqQcS3QAAAAAAAADQRp555hntvPPOmj59eu6m1EZP7gYAAAAAAAAAAOaaOnUqSe6FRKIbAAAAAAAAANrIUkstlbsJtUPpEgAAAAwKG0ICAAAAC2f69Ok68cQTFRFzvpZSmvP3f6rPjVcj0Q0AAIBBGTt2bMtj9vX1SZLGjx+fJS4AAAAwGGeffTYTRipGohsAAAC11Nvby8MBAAAAamnPPffUe9/73jkzusszuyXpb3/7my6//PIcTastEt0AAAAAAAAA0ELDhg17zZWR48aN05prrqkHH3zwVUnwhnPPPbdJrasnEt0AAAAAAAAA0EZmzZqlww8/PHczaqUndwMAAAAAAAAAAHMdeOCBuZtQOyS6AQAAAAAAAKCNfP3rX8/dhNqhdAkAAAAAAAAAtJExY8ZowoQJr/kz48aNa1Fr6oEZ3QAAAAAAAACAWmNGNwAAAGqpv79f48ePb2nMvr6+lsYDAABAd3r66af1mc98RjNnzszdlNpgRjcAAAAAAAAAtJHbb7+dJPdCYkY3AAAAaqm3t1cTJ07M3QwAAACgcrNnz87dhNphRjcAAAAAAAAAtJFFFlkkdxNqhxndAAAAqCVqdAMAAKBTvfzyy7mbUDvM6AYAAAAAAACANpJSyt2E2iHRDQAAAAAAAABtZNlll83dhNoh0Q0AAAAAAAAAbeS5557L3YTaoUY3AAAAaqm3t1cTJ07M3QwAAACgcmussYY22mgj3Xnnnerp6VFKac4fyaVNHnjggbyNbDMkugEAAAAAAACgjTz//PO67rrrcjejVihdAgAAAAAAAABt5MUXX8zdhNphRjcAAEBFKKUBAAAAoAojR47UmDFjNG3atNxNqQ0S3QAAABXp7+/X+PHjWxqzr6+vpfHmheQ+AAAAUK0nnniCJPdCItENAACAQRk7dmzLY/b19am3t7flcQEAAIBWWHHFFbXddttpypQpr9iAsvz3zTffnK197YhENwAAAGqpW2fQAwAAoPNNmTJFF198ce5m1AqbUQIAAAAAAABAG3nsscdyN6F2mNENAACAWmLzTwAAAHSqxRdfPHcTaodENwAAAGqJ0iUAAADoVKuuumruJtQOiW4AAICKMMO4tXi9AQAA0Kn++c9/5m5C7ZDoBgAAqAgzjLsHCXYAAAA004gRI3I3oXZIdAMAAGBQujHpO3bs2JbHZFADAACge7zwwgu5m1A7JLoBAAAwKDmTvsygBwAAQCdadNFFczehdkh0AwAAoJao0Q0AAIBOde+99+ZuQu2Q6AYAAEAtURMdAAAAner666/P3YTaIdENAACAWmJGNwAAADrVZpttpj/84Q+5m1ErPbkbAAAAAAAAAACYa/XVV8/dhNoh0Q0AAAAAAAAAbeShhx7K3YTaoXQJAAAAaoka3QAAAOhUN910U+4m1A6JbgAAANQSNboBAADQqdZdd11dcMEFuZtRKyS6AQAAUEvM6AYAAECnmjx5cu4m1A41ugEAAAAAAACgjXzsYx/L3YTaYUY3AAAAaonSJQAAAOhUM2bMyN2E2iHRDQAAgFqidAkAAAA61QorrJC7CbVD6RIAAAAAAAAAaCPPPvts7ibUDjO6AQAAAAAAAKCNjBkzRhMmTHjNnxk3blyLWlMPJLoBAABQS9ToBgAAQF0ddthhuvbaa3M3o6NQugQAAAAAAAAAWuiRRx7J3YSOw4xuAAAA1BKbUQIAAKCujj76aJ100kmKCEWEJM35W5LuvfdeTZkyJVfzaolENwAAAAAAAAC00I9+9CNdddVVuZvRUUh0AwAAoJao0Q0AAIC62m233TRixAhJUkpJKaU530spadq0abrhhhtyNa+WSHQDAACglihdAgAAgLpafPHF9dWvfnW+37/jjjtIdC8kEt0AAACoJWZ0AwAAoFNNnz49dxNqpyd3AwAAAAAAAAAAcz366KO5m1A7zOgGAABALVG6BAAAAJ1qyy231Lnnnqunn346d1NqgxndAAAAAAAAANBGhg4dqpEjR+ZuRq2Q6AYAAAAAAACANrLHHnto6tSpuZtRK5QuAQAAAAAAAIAWmjRpkg488MDczegoJLoBAAAq0tvbq4kTJ+ZuRtfg9QYAAEBdTZ8+PXcTOg6JbgAAgIqwOWJr8XoDAACgrrbaaitttdVW8/3+o48+qu23376FLao/anQDAAAAAAAAQBt5/vnnczehdkh0AwAAAAAAAEAbefLJJ3M3oXYoXQIAAAAAAAAAbWSdddbRJptsoltvvVWSlFJ61c888cQTrW5WWyPRDQAAAAAAAABt5K677tK1116buxm1QukSAAAAAAAAAGgjM2bMyN2E2mFGNwAAAAAAAAC00EUXXaRTTz01dzM6ColuAAAA1FJvb68mTpyYuxkAAADAQnvrW9+auwkdh0Q3AAAAaqm/v1/jx49vacy+vr6WxgMAAEBniojcTeg41OgGAAAAAAAAgBa67777cjeh4zCjGwAAAAAAAABaaNttt9W222473+/fdtttLV+9WHckugEAAFBL1OgGAABAp5o5c2buJtQOiW4AAADUEjW6AQAA0Kmo4b3wqNENAAAAAAAAAG1k9OjRuZtQOyS6AQAAAAAAAKCNTJs2LXcTaodENwAAAAAAAAC0kWuvvTZ3E2qHRDcAAAAAAAAAtJEddtghdxNqh80oAQAAUEu9vb2aOHFi7mYAAAAAlVt99dW11157aerUqZK8OWV5g8qI0JVXXpmreW2JRDcAAAAAAAAAtJGrr75axx9/fO5m1AqJbgAAANRSf3+/xo8f39KYfX19LY0HAACA7rTaaqvlbkLtkOgGAABALVG6BAAAAJ3qjjvuyN2E2iHRDQAAgFpiRjcAAAA61bBhw3I3oXZIdAMAAKCWcs7oZiY5AAAAmmnMmDG5m1A7JLoBAABQSzlndLc6bjk2AAAAOt+iiy6auwm1Q6IbAAAAtUSNbgAAAHSqIUOG5G5C7ZDoBgAAAAAAAIA28pa3vEVHHXWUHn74YUlSSmnOn8bnxxxzTM4mth0S3QAAAAAAAADQRm655Rbtu+++uZtRKyS6AQAAUEs5a3QDAAAAzTRt2rTcTagdEt1YIN1YAzPnMff392eJCwBAneS8V3dbvwgAAACtNXv27NxNqB0S3Vgg3ThjKscxS/mPGwCAusjZP6GPAAAAgGbaeuut9eKLL+q+++5TRCgi5nyv8fkNN9yQsYXth0Q3AAAAaqkbV5wBAACgO/T09GjUqFF6/vnnJekVm1Bi3kh0AwAAAAAAAEAbuf3223X00UfnbkatkOgGAAAAAAAAgDbyjne8QwcffLAefvjheZYtkaQzzzwzV/PaEoluAAAA1FI37iECAACA7pBS0gc+8AHNnDlznmVLUkokugcg0Q0AAAAAAAAAbeSee+7RrrvumrsZtUKiGwAAALXEZpQAAADoVM8++2zuJtQOiW4AAADUEqVLAAAA0KnWXntt7bzzzrrvvvteUbKkIaWk6667LkPL2heJbiyQbpwxlfOY+/v7s8QFAAALptv6RQAAAGitIUOG6Atf+MJr/gyJ7lci0Y0F0o0zpnIcs5T/uAEAqIucg9Jjx45teUz6CAAAAN1j+vTp+tSnPpW7GbVCohsAAAC11I0D8QAAAOgOL774Yu4m1A6JbiwQSpe0FqVLAAD4z7qxfwIAAIDuMHr0aE2YMOE1f2bcuHEtak09kOjGAunGGVOULgEAoL11Y/8EAAAAwLyR6AYAAEAtMaMbAAAAnSoidP755+uOO+5QT0+PUkpKKUnSKz7GXCS6AQAAMCg5S30xoxsAAACd6IgjjtBvf/vb3M2oFRLdAAAAGJSxY8e2PCYJZwAAAHSycePGkeheSCS6AQAAUEuULgEAAEBd3XTTTdpvv/1yN6OjkOgGAACoCIlXAAAAAAvipptuyt2EjkOiGwAAoCLdWjOa5D4AAACwcC677LLcTeg4JLoBAAAwKDlrdHfjwAIAAADq77LLLtOXv/xlSVJK6VXff+qpp1rdpNoj0Q0AAAAAAAAALTRq1Chdfvnl8/3+5ZdfrhNOOKGFLao/Et0AAACoJWqiAwAAoFNttdVWuu666/S3v/0td1Nqg0Q3AAAAaqlba6IDAACg802cOJEk90Ii0Y0F0o0zpnIec39/f5a4AABgwXRbvwgAAACt1dvbm7sJtUOiGwukG2dM5ThmKf9xAwBQFzkHpXNuwAkAAIDO19PTo56eHs2ePTt3U2qDRDcAAABqqRsH4gEAANAdXnjhBZLcC4lENwAAQEW6sdRXTrzeAAAA6FQPPPBA7ibUDoluLJBufJCkRjcAAO0t54zubusXAQAAoLUmTZqUuwm1Q6IbC6QblwZToxsAsLC68X4p5U36UqMbAAAAnWibbbbRNddck7sZtUKiGwAAAIOSK+nLTvQAAADoVMsvv7yWWGIJPffcc7mbUhskugEAAFBL3TqDHgAAAJ3vySefJMm9kEh0AwAAAAAAAEAbWXnllbX//vvrwQcflCSllF71M+ecc06rm9XWSHQDAAAAAAAAQBv505/+pMMPPzx3M2qFRDcAAABqqbe3N+tGmAAAAECzPPbYY7mbUDskugEAAAAAAACgjXziE5/QqFGj9MgjjygiJEmzZ89WRMz5c/rpp2duZXsh0Q0AAIBaYjNKAAAAdKqXX35ZZ5xxhu6///7cTakNEt0AAACoJUqXAAAAoFNdccUVJLkXEoluLJBufJDMecz9/f1Z4gIAAAAAACC/rbfeWvfee69uu+22OaVLyn9HhB599NGcTWw7JLqxQLpxaXCOY5byHzcAAAAAAADy6unp0f+3d+9RepX1vcB/e4ZLwrXcllpU0IohwctbewSyWB5nuHgsshSPiosFKCpFUMBx0UKry3Nsi0tEK+M1Hgo2HKwKB7oknmgthYk9lUgK+BYhoZJyy0UuJjEXSAaZ2ecPchtgwgzzZj/vs/fns9aszOz95n2+e/iDN9/39z77z/7sz3b4mP7+/orS5EHRDQBAlpr4RjwAAM3w5JNPxplnnhmPP/546ijZ6EkdAAAAAACAba6++mol9yQpugEAAAAAushRRx2VOkJ2bF0CAECWmnizbAAAmmFkZCR1hOwougEAOkTxCgAAdMIee+yROkJ2FN1MSBP/4Z7ymtvtdpJ1AZgaN0esVsrfd9NeFwEAUK3e3t7UEbKj6GZCmvgP9xTXHJH+ugEgFynflO7r66t8Ta8RAACaY3h4OHWE7LgZJQAAAABAF3n5y1+eOkJ2THQDAJClJn7iDACAZli9enXqCNkx0Q0AAAAA0EV23XXX1BGyY6IbAAAAAKBDRkZG4vjjj08do3FMdAMAAAAAdEhPT08cddRRqWM0joluAACy1Gq1YsGCBaljAADAGEVRxKWXXjql52i32/HJT36yQ4maQdENAECW3IwSAICcjYyMjHvOHt2Tp+gGACBLJroBAMjVVVddFd/5zndSx6gVe3QDAAAAAFToLW95S/T09Iz7xeSZ6AYAAAAAqNBrX/vauPnmm8c9f91118WcOXMqTJQ/RTcAAEySLVMAANiZZs2alTpCdhTdAABkKeXNKKted/u1AQCoPzejnDxFNwAAWXIzSgAA6mrhwoWpI2THzuYAAAAAAF3kjW98Y+oI2THRDQBAllJuXQIAADvT9OnTU0fIjqIbAAAAAKBCP/nJT+LSSy9NHaNWbF0CAAAAAFAhN5vsPBPdAAAAAAAV6u/vj6OPPnrMsbIst35/0003xVe+8pWqY2VN0Q0AQJZarVYsWLAgdQwAAJi0K6+8Mr773e+mjlErti4BAAAAAKjQscceG3vssUdMnz79eb+YPBPdAABkqd1ux8DAQKVrDg4OVroeAAD19OpXv3rMRPf225ZERAwPD8epp576nOOMT9ENAECWbF0CAECurrjiivj+97+fOkatKLoBAAAAACr0tre9LX70ox8977miKGLjxo3x1FNPVZwqb4puAACyZOsSAABy9apXvSpuvPHGcc/ff//98ZGPfKTCRPlzM0oAAAAAgC7ihpSTp+gGAAAAAOgijz/+eOoI2VF0AwAAAAB0kVe84hXR06O6nQx7dAMAkKVWqxULFixIHQMAADrurrvuitHR0dQxsuJtAQAAAACALnL33XenjpAdE90AAGSp3W7HwMBApWsODg5Wuh4AAM10zDHHxPXXX586RlZMdAMAAAAAdJE99tgjdYTsmOhmQpq4B2bKa26320nWBQAmpmmviwAAqJYbUU6eopsJaeJHg1Ncc0T66wYAXlhfX1/la3qNAADQHNOnT08dITveGgAAAAAA6CL/8i//kjpCdkx0AwCQpSZurQYAQDMMDw+njpAdRTcAAAAAQIWGhobir/7qr1LHqBVFNwAAWWriPUQAAKgHE9udp+gGAAAAAKjQ29/+9nj729++w8esXbs2Nm3atPXnsizHnD/11FN3SrZcKboBAMiSPboBAKirtWvXxllnnRW/+c1vUkfJRk/qAAAAAAAAbLN27Vol9ySZ6AYAIEv26AYAoK6evU0JL0zRDQBAlmxdAgBAXd17772pI2TH1iUAAAAAAF1k5syZqSNkR9ENAAAAANBFNmzYkDpCdmxdAgAAk2TLFAAAdqaDDjoodYTsKLoBAGCS+vr6Kl/TjTABAJpjxYoVqSNkR9ENAECW2u12DAwMVLqmshkAgCrsu+++qSNkxx7dAAAAAABd5J/+6Z9SR8iOiW4mpNVqNW4vypTX3G63k6wLAExM014XAQDQWXPnzo2rr746dYxaUXQzIU38aHCKa45If90AwAuzRzcAAFNx9NFHK7o7TNENAECWmviJMwAA6uHwww+PoaGhcc/fc889cd5551WYKH+KbgAApkTZDAAAnTVt2rTUEbKj6AYAYEpSbePRarUqXxcAAKqwfPny1BGyo+gGAAAAAKhQWZaxatWqiIgoimLMnxER++yzT5JcOVN0AwCQpSbeLBsAgHo444wzYsWKFalj1EpP6gAAAAAAAE0ye/bs1BFqx0Q3AAAAAEAHLVq0KG677batP5dl+ZzHvPvd7x737//qV7+Ke+65Z6dkqytFNwAAAABAh4yMjMTFF1+cOkbjKLoBAMhSq9WKBQsWpI4BAABj9Pb2xre+9a345S9/Oeb49jeb3P7Y9tPeWx7z2GOPxbXXXrtzg9aMohsAoEMUrwAAQETEjBkzYsaMGS/67990000dTNMMim4AgA5pt9sxMDBQ6ZqDg4OVrgcAAOx8a9asSR0hOz2pAwAAAAAAsM3xxx//vFudMD4T3QAAZMkEPQAAdbX//vvHLbfcssPH9Pf3V5QmDya6AQAAAADImoluAACy5OafAADAFopuAACyZOsSAADq7NZbb42VK1emjpENRTcAAAAAQBe555574tOf/nTqGFlRdAMAMCUptw+xdQkAAHU0a9asGBgYiIcffnjcx/zDP/xDhYm6n6IbAIAp6evrq3zNwcHBaLVala8LAABV+M1vfmPbvEnqSR0AAAAAAIBtHnvssdQRsmOiGwCAKbF9CAAAdNbq1atTR8iOohsAgClJtXVJRMTAwECSdQEAYGc66qij4qijjoo77rhjzPGiKLZ+/7vf/a7qWF1N0Q0AAAAA0EUefPDBuO2221LHyIqimwlptVqN+1hyymtut9tJ1gWAnDTx9QkAAM2w1157pY6QHUU3E9Jutxv30eAU1xyR/roBAAAASGv7LUqYmJ7UAQAAAAAA2Mb+25NnohsAoENspQEAAHTCI488kjpCdhTdAAAd0sStvlLy+wYAoK7e9KY3xfve975YunRp9PT0jPkqiiJ6enriX//1X1PH7CqKbgAAAACALrLLLrvExz72sR0+pr+/v6I0eVB0AwB0iK1LquX3DQAAbKHoBgDokKZupaFsBgAAUlN0AwB0iAljAACANBTdAAAd0tSJ7r6+vsrXHBwcjFarVfm6AABAd1J0AwAAAABU6M4774wLL7wwdYxaUXQDAJClpk7QAwCQv1tuuSV1hNrpSR0AAAAAAKBJ9thjj9QRasdENwAAWXLzTwAAcnXuuefGySefHEVRRFmWERFb/yyKIubNmxfXXnttyojZUXQDAAAAAHTQyMhIPPXUU897bku5vd9++0VEPKfoLssyTj/99DjwwANj2bJlz/n7ZVnG6OhozJ8/fyelz5OiGwCALNmjGwCAbjQyMhLHH3986hiNo+gGACBLti4BAKAb9fT0xPvf//648cYbtx7bMq09UcPDw52OVXuKbgAAAACADimKIs4555w455xzXvRz3HXXXfGJT3yig6nqT9ENANAhTZ0wTnXNti4BAKCufvvb36aOkB1FNwBAhzS1eO3r66t8zcHBwca+sQAAQP297nWvSx0hOz2pAwAAAAAAsM2qVatSR8iOiW4AALLU1Al6AADqb8WKFakjZMdENwAAAABAF5k5c2bqCNkx0c2ENHEPzJTX3G63k6wLADlJ+f/qpr0uAgCgWo899ljqCNlRdMM4UnwcOsJHogEgB6luwAkAQDOsX78+dYTsKLqZEHtgAgDdxusTAADq6pBDDkkdITuKbgAApiTlNh62EAEAoI4OPvjgGBoa2uFj+vv7K0qTB0U3AABTknIbDxPdAADU0erVq+M973lP6hhZ6UkdAAAAAACAbe6+++7UEbJjohsAgCy1Wi1blwAAUEt77rln6gjZUXQDAJAlN6MEAKCu/uiP/ii++MUvxiOPPBJlWUZZljE6Orr1+7Is4+tf/3rqmF1F0Q0A0CEmjKvl9w0AQK6+973vxRVXXJE6Rq0ougEAOsSEcbX8vgEAyNXMmTNTR6gdRTcAAAAAQIVarVYMDQ2Ne37VqlXx3ve+t8JE+VN0AwCQJVuXAABQVwcccMAOi/CIiP7+/orS5EHRDQAAAADQRZYtWxYf+MAHUsfIiqIbAIAs2aMbAIC62meffWLPPfeMJ554InWUbCi6AQDIkq1LAACom7IsIyJi7733jnnz5o059uzHnXDCCZVm63aKbiakif+QTHnN7XY7yboAAAAA7Hzf/va345prrkkdo1YU3UxIEz8anOKaI9JfNwAvXhPfGAYAACbvmGOOUXR3mKIbAKBDmvjGcEp+3wAA5GrGjBkxNDQ07vn169fH6aefHuvWraswVd56UgcAAAAAAGCb2267Tck9SSa6AQDIkq1iAACoq9e85jWpI2RH0Q0AQJZsXQIAQF0tX748dYTsKLoBAMiSiW4AAOrqDW94Q+oI2VF0AwCQJRPdAADU1apVq1JHyI6iGwAAAACgQk8++WTccMMNUZbl857/0Y9+VHGi/Cm6AQAAAAAq9M1vfjPmz5+fOkatKLoBAAAAACp07rnnxv777z/m2PbT3T/72c/igQceqDpW1hTdAAAAAAAV2nPPPePDH/7wuOdPOeWU+NjHPhbLly+vMFXeFN1MSKvVigULFqSOUamU19xut5OsCwBMTNNeFwEAUK2nn35ayT1Jim4mpN1ux8DAQKVrDg4OVrres6W45oj01w0AuUj5pnRfX1/la3qNAADQHAYrJk/RDQAAk+QfHgAA7Ez77rtv6gjZUXQDAJCllJ8486kvAAB2pqOOOir23XffWLt2beoo2VB0MyH26K6WPboBoLs17XURAADVWr58uZJ7khTdTIg9uquT+roBIBf26AYAoK4OOeSQ1BGyo+gGACBLTXwjHgCAZrj//vtTR8hOT+oAAAAAAABss+eee6aOkB1FNwAAAABAF9lvv/3i93//91PHyIqtSwAAyFITb5YNAEAz7LbbbnHooYfGunXroizLMV8REWVZxvDwcOKU3UXRDQBAluzRDQBAXd1yyy1x6623po6RFUU3AABZMtENAEBdvexlL0sdITuKbgAAsmSiGwCAuvrDP/zDuPrqq2PdunVRFMXWr56eZ2652NPTE2effXbilN1F0Q0AQJZMdAMAUGevfOUrU0fIiqIbAIAsmegGAKDOli5dGqtWrYqIGHMjyoiI0dHRVLG6lqIbAAAAAKCLzJ8/P770pS+ljpEVRTcAAFmydQkAAHV1xBFHpI6QnZ7UAQAAAAAA2ObRRx9NHSE7JroBAJgSU9UArTZcOAAAF/hJREFUANBZM2fOjJe85CUK70lQdAMAMCV9fX2Vr7nlppBuRgkAQB2tX79eyT1Jti4BAAAAAOgi06dPTx0hO4puAAAAAIAu8tRTT6WOkB1FNwAAAABAF7nssstSR8iOPboBADqk1Wq5MSMAADBlH/jAB+IXv/hF6hhZUXQDAHRIu912c0QAAGDKXv/618eHPvSheOCBB6Isy+d9zE9/+tOKU3U3RTcAAFOScordBD0AAHXU29sb73vf++Lpp58eU3Rv/72ieyxFNwAAU9LX11f5moODg9FqtSpfFwAAqrB06dL4kz/5k9QxsqLoBgAgS7aKAQCgrg4++OA4+uij4+67746IiKIotp7b8v26deuSZOtWim4AAAAAgC4yffr0+PznP7/Dx/T391eUJg+KbgAAAACALrJp06a49NJL45577nne89tPePMMRTcAAAAAQBdZtmyZm01OkqIbAAAAAKCLHHbYYXHjjTfG8PDwc86VZRkREe9///urjtXVFN0AAAAAAF1mn332SR0hK4puAACmZMGCBY1cGwAAdpZNmzbFe9/73njiiSdSR8mGohsAgCnp6+urfM3BwcFotVqVrwsAAJ2wdOnSuOCCC7ZuQ/JsmzZtqjhR/hTdAABkqd1ux8DAQKVrDg4OVroeAAD1NDQ0FBs3bkwdo1YU3QAAZKnVatm6BACALJ111llx+umnj3v+rrvuij//8z+vMFH+FN0AAAAAABUqiiKmT58+7vk99tijwjT1oOhmQpo4MZXymtvtdpJ1AYCJadrrIgAAqnXnnXemjpAdRTcT0sQ9MFNcc0T66wYAXliqG3ACANAMu+yitp0svzEAALLUxDfiAQCoh3Xr1sVVV10VERFlWT7n/E9+8pOqI2VP0Q0AQJaauLUaAAD1cMUVV8T8+fNTx6gVRTcAAFky0Q0AQK7OP//8OOyww55zvCiKiHhmj+6f/vSnVcfKmqIbAAAAAKBCu+++e7zrXe8a9/yMGTMU3ZOk6AYAYEpSbh9i6xIAAOpoxowZ8fd///exfv36Mce338/73HPPrTpWV1N0AwAwJX19fZWvuWULEVuXAABQR2vWrInTTjstdYys9KQOAAAAAADANnfddVfqCNkx0Q0AAAAA0EWOPvroeOtb3xqLFy+OiG03qdzyZ0TEo48+miRbt1J0AwAAAAB0kd133z1OP/30WLVqVUSM3Zt7i7/4i7+oOlZXU3QDAJClVqvlZpQAANTSwoUL41Of+lTqGFlRdAMAAAAAVGjDhg1xzTXXxOjoaJRluXVie8uf999/f8p4WVJ0AwAAAABU6Fvf+lbMnz8/dYxaUXQzIU38aHDKa26320nWBQAmpmmviwAA6Ky99tordYTaUXQzIe12OwYGBipdc3BwsNL1ni3FNUekv24A4IX19fVVvqbXCAAA9dHX1xfXXntt6hi1ouhmQkx0V8tENwC8sJRvxDftdREAAJ11+OGHx9DQ0Ljn//Zv/za++93vVpgof4puJsREd3VSXzcA8MJMdAMAsDOdeOKJiu5JUnQDAJClJn7iDACAZjjwwAPjuOOOiyVLloz7mJUrV1aYqPspugEAAAAAushDDz0UN998c+oYWVF0AwCQpSZurQYAQDOUZZk6QnYU3QAAZMnWJQAA5OrOO++MCy+8MHWMWlF0wzhS/uO53W4nWRcAXgxlMwAATM7SpUtTR6gdRTeMI8XHoSN8JBqA/PT19VW+5uDgYLRarcrXBQCATjjllFPilFNOGff86tWr4z3veU+FifLXkzoAAAAAAADb3H///akjZMdENwAAWXIzSgAA6uquu+5KHSE7im4AAKYk5R7d9gcHACBHt956a3z6059OHaNWFN1MSMobM6biZpQAMDGp9uiOiGQT3U17XQQAQGcpuTtP0c2ENPGjwW5GCQCMJ2W5DwBA/ubNmxdz586NiIiiKMb8ueX79evXx8aNG7f+XBRFlGUZZVlGURSGL55F0Q0AQJaa+IkzAADqYe+9947zzz9/3PP33XdfnH322RUmyp+iGwCALDXxE2cAADRDWZapI2RH0Q0AQJZMdAMAUFcbNmxIHSE7im4AgA5RvFbLRDcAAHV1xBFHxLHHHhv33nvvuI9ZuXJlhYm6n6IbAKBDFK8AAEAnPPzww3HLLbekjpGVntQBAAAAAADYZv369akjZMdENwAAWbJVDAAAdXXrrbemjpAdRTcAAFmyVQwAAHX1x3/8x3HDDTekjpEVRTcT0sSJqZTX3G63k6wLAExM014XAQBQreHh4dQRsqPoZkKaODGV4poj0l83APDC+vr6Kl/TawQAgObYZRe17WT5jQEAMCUpp5tNVgMAABGKbgAApijVdHOr1ap8XQAA6ITh4eH48Y9/HGVZPu/5DRs2VJwof4puAAAAAIAKfe1rX4v58+enjlErim4AALLUxHuIAABQD2efffa4+3AXRRE/+MEPKk6UP0U3AABZarVa9ugGACBL++yzzw6HNvr6+iof6sidohsAgCyZ6AYAoK7e+MY3xtDQ0A4f09/fX1GaPPSkDgAAAAAAAFNhohsAAAAAoIsMDw/HpZdeGosXL47R0dGIiBgdHY2yLLd+MZaiGwCALNmjGwCAuvrZz37mte4kKboBADpE8QoAAHTCoYcemjpCdhTdAAAd4uaIAABAJ6xbty51hOwougEAyJI3FgAA6Eajo6Nx3HHHpY7ROD2pAwAAAAAA1MmsWbNSR2gcE90AAGTJnugAAHSjnp6e+MY3vrHT1+nv79/pa+TERDcAAAAAAFkz0Q0AAAAA0EUeeuihOPPMM1PHyIqiGwCALLkZJQAAdbVp06bUEbKj6AYAIEv26AYAoK4OOuig1BGyo+gGAAAAAOgi+++/fwwNDe3wMW5GOZaiGwCALNm6BAAA2ELRDQAAAADQISMjI7Fo0aK477774rDDDosjjzwyent7J/Uca9eujdNOOy2eeOKJnZSyfhTdAAAAAECjFEUxMyI+EREHRsTNEfHOiHgoIjaUZfmnL/Z5R0ZG4qKLLoolS5bEpk2bYtq0aTFz5sy47LLLJlV2r169Wsk9SYpuJqSJN3tKec3tdjvJugCQk5T/r27a6yIAgLopy3JJRJxTFEVPRPzviHgyInoi4tGpPO+iRYtiyZIlsXHjxoiI2LhxY/z7v/97XHzxxfHSl7506+MWLFigyO4wRTcT0sQ9MFNcc0T66waAXKR8feI1AgBA/oqieGdEnBsR10TE98uyHC2K4stFUbyhLMsX9Zz33XdfbNq0acyxkZGRuOOOO6aclx1TdAMAAAAAjVOW5byImFcUxfyyLL+7+fBjEbHXi33Oww47LKZNm7Z1ojsiYtddd413vOMd8Qd/8Adb1o0bbrghHnrooSmk59kU3QAAZKmJW6sBANAZRVH0RcR/j4jdI+JHRVFcHc9sX7JLRFz2Yp/3yCOPjJkzZ8bixYtjeHg4dt9995g1a1acd955Y/bofulLXxoXXXTR1C6CMRTdAABkqYlbqwEA0BllWS6IiAXbHfpGJ563t7c3Lrvssli0aFEsXbo0XvOa18SRRx75nBtRvvnNb46hoaFxn2fevHlx+eWXdyJSYyi6AQDIkoluAAC6UW9vb8yePTtmz579op/jxBNPjCVLlsQ999wTPT090dPTE0VRRG9vbxRFET09PXHvvfd2MHX+FN0AAGTJRDcAAHW1yy67xMUXX7zDx/T391eUJg+KbgAAsmSiGwAA2ELRDQBAlkx0AwAAWyi6AQAAAAC6yIoVK+L0009PHSMrPakDAAAAAACwze233546QnYU3QAAAAAAXeS+++5LHSE7ti4BAAAAAKjQ0qVL4/zzz4+RkZGIiCjLcutXRMTo6GjKeFlSdAMAkKVWqxULFixIHQMAACZtaGgoNm3alDpGrSi6AQAAAAAqdNZZZ8UZZ5wRRVFEURQREWO+X7BgQXzuc59LGTE7im4AAAAAgAoVRRHTpk0b9/zxxx8fb33rW2N0dHTrdiZbbPn5xBNP3KkZc6PoBgAgS+12OwYGBipdc3BwsNL1AABorl133TV1hKwougEAyJI9ugEAqKsHH3wwPvShD6WOkRVFNwAAWTLRDQBAXR144IFxyCGHxEMPPZQ6SjYU3QAAAAAAXWSvvfaKuXPn7vAx/f391YTJhKIbAAAAAKBCv/vd72LhwoVjbjS55fuyLOPJJ5+ML33pS6niZUnRDQBAluzRDQBArr72ta/FD3/4w9QxakXRDQAAAABQoQ9+8IOxdu3aKMsyiqLYenz775/v2PbfDw0N7dyQmVF0AwCQJTejBAAgVwcccED85V/+5bjnH3nkkTj11FMrTJQ/RTcAAFmydQkAALlas2ZNzJkzZ8we3RHb9uletmxZilhZU3QDAHSI4rVaJroBAMjVVVddFTfddFPqGLWi6AYA6BDFKwAAMBEXXHBBvOENb4iI59+X+84774x//Md/rDpW1hTdAABMScopdhP0AADkaLfddou3ve1t457/7W9/W2GaelB0AwAwJX19fZWvuWWS3QQ9AAB19PrXvz51hOwougEAyJI90QEAqKunnnoqdYTsKLoBAMiSPdEBAKir3XbbLXWE7Ci6mZAmTkylvOZ2u51kXQBgYpr2uggAgGpNmzYtdYTs9KQOAAAAAADANo8++mjqCNkx0c2ENPGjwSmuOSL9dQMALyzlDTgBAKi/J554InWE7Ci6AQDIUhO3VgMAoB5GR0dj2bJlUZZllGUZETHmz//8z/9MGS9Lim4AAAAAgArNmTMnrr/++tQxakXRDQAAAABQoXe+853xb//2b2OOFUURRVFERMQDDzyQIlbWFN0AAAAAABV6xSteEXPnzh33/M033xyXXHJJdYFqQNENAECWmnizbAAAmuG4446Lww47LNauXTvm+JaJ74iI888/v+pYXU3RDQBAltyMEgCAOnvlK1+ZOkJWFN0AAB2ieG0O/50BANiZHnnkkTj11FNTx8iKohsAoENspVGtlL/vqtfdfm0AAOqvt7c3dYTsKLoBADrERHe1/L4BAMjVD37wg/jKV76SOkatKLoBADrERDcAADARL3nJS1JHqB1FNwBAh5gwBgAAJmL27NkxNDQ07vmNGzfGJZdcEr/85S+jLMsoyzIiYuv3ZVnGpk2bqoqbBUU3AECHmOiult83AAC5mjdvXlx++eWpY9RKT+oAAAAAAABNcsABB6SOUDuKbgAAAACACm3cuDF1hNqxdQkAAFmyJzoAALnasuc2naPoBgAgS/boBgAgVyeccEKccMIJO3zM4sWLY9WqVVt/LopizPnPfOYzOyVbrhTdAABkyUQ3AAB1tXz58vj4xz+eOkZWFN0AAGTJRDcAAHW1cuXK1BGy42aUAAAAAABdZO+9904dITsmugEAyFLKrUtsmQIAwM60cOHC1BGyo+hmQpq4B2bKa26320nWBQAAACC9u+++O3WE7Ci6mZAm7oGZ4poj0l83AExW094Mj4jo6+urfE2vEQAAmuOcc86Jj370o6ljZEXRDQDAlKQqfVutVuXrAgBAFXbbbbfUEbKj6AYAIEtN/MQZAADNcOihh8bQ0NAOH9Pf319Rmjz0pA4AAAAAAABTYaIbAKBDmnjz5oi0e3Q38fcNAAA8l6IbAKBDmrqVhj26AQCgsx5++OH44Ac/mDpGVhTdTEgTJ9RSXnO73U6yLgDkJOUbC017XQQAQLV+7/d+Lw466KB4/PHHU0fJhqKbCWnihFqKa45If90AwAtLNcUOAEA9XH311TF37tzUMWrFzSgBAAAAACr05je/OXWE2jHRDQDAlLgZJQAATM7Pf/7z1BFqR9ENAMCUpNzGo2lbqwEAUA/HHHNMXHPNNalj1IqiGwCAKTHRDQAAkzNjxowYGhoa9/zixYvj4x//eIWJ8qfoBgBgSkx0AwBAZ82aNSv+5m/+Jn79619HRERRFFvPFUURRVHEF77whVTxupKiGwAAAACgiyxdujQuvPDC1DGyougGAAAAAKjQwoUL41Of+lTqGLXSkzoAAAAAAECTPProo6kj1I6JbgAAstRqtdyMEgCALJ188slx8sknj3v+uuuuizlz5lSYKH8mugEAAAAAusj++++fOkJ2FN0AAAAAAF1kl11sxDFZim4AAAAAgC4yffr01BGyo+gGAAAAAOgiS5YsSR0hO2bgAQAAAAAq9MMf/jC+/OUvp45RKya6AQAAAAAqpOTuPBPdAABkqd1ux8DAQKVrDg4OVroeAAD19M///M9x3333bf25LMut3xdFEaOjo3HdddfFgw8+uPVcWZZbvyIiVqxYUW3oLqfoBgAAAACoUG9vbxx++OE7fMxnP/vZHZ7v7+/vYKL8KboBAAAAALrI6OhofO9734sHHnhgzLQ341N0AwAAAAB0kSVLlsSVV16ZOkZWFN0AAEzJggULGrk2AADsLMuXL08dITuKbgAApqSvr6/yNQcHB6PValW+LgAAdMKaNWtizpw5Y44VRbH1T0X35Cm6AQDIUrvdjoGBgUrXHBwcrHQ9AADq6atf/apPJ3ZYT+oAAAAAAABNsuuuu6aOUDsmugEAAAAAKvTJT34yXv3qV497/uGHH44f//jHFSbKn6IbAAAAAKBCGzZsiDvuuCPKsoyIbftzb3H77beniJU1RTcT0mq1GrdvUMprbrfbSdYFYGqa+P/LpvLfGQCAqfjOd76jzO4wRTcT0sSbPaW45oj01w0Ak5Wy9E21dl9fX+Vreo0AAFAfF1xwQbzlLW+JiNg61b29RYsWxfXXX191rKwpupmQJk6omegGYLKa+MZwRNrSN9Xvu2mviwAA6KyiKOKAAw7Y+vOWsrssyyjLMu6+++5U0bKl6GZCmvgPdxPdANDdUr4pbaIbAICp+OY3vxk33HBD6hi1ougGAAAAAKjQySefvPVmlFtuRLn9n2vWrIk1a9akjJgdRTcAAFlq4ifOAACoh5e//OXxd3/3d+OeX7lyZZx22mkVJsrfhIvuoih6I+L2iFhRluVJ2x3/akR8uCzLvTb/fGZEfDEiVmx+yNfLsrxy87mRiPjl5uMPl2X5zs3HXxUR34+IAyLijog4oyzLp6ZwXQAA1FwT7yECAEA9zJs3Ly6//PLUMWplMhPdn4iIJRGxz5YDRVH8l4jY73kee21Zluc9z/GNZVm2nuf4FyLi8rIsv18Uxbci4iMRMWcS2QAAaBgT3QAA5GrJkiWpI0zKwoUL48gjj4ze3t4X/RxFUewZET+NiM9GxOyIeF9EvKksyw2dyNgzwRAvj4h3RMSV2x3rjWcmty+aSoDimc1njo2I6zcfujoiTp7KcwIAAAAAdKsjjjgidYRJ+eu//uu46KKLYmRkZCpPc3FEXBcRUZblpyPi1k5k22KiE92D8Uyhvfd2x86LiHllWf56y0bp23lPURT/NSJ+FRGfLMty2ebj04qiuD0ino6IS8uy/EE8s13Jb8uyfHrzY5ZHxMGTvxQAAAAAgO530kknxUknnTTu+f/4j/+Ic845p8JEO7Zx48ZYvHhxLFq0KGbPnj3pv18UxQkRsTgipnU83JY1yrJ8oRAnRcSJZVl+rCiKvoj404g4O55p3/vKsny6KIoN2+3RfUBEbCjLcrgoio9GxPvLsjx287mDy7JcURTFqyPilog4LiLWRsTPy7J8zebHvCIiflyW5et2xgUDAAAAADC+/v7+z8QzW4xsvyPIaET8z6GhoUsm+3xFUXwuIvaMiFkRsTEi3h0R346I8zq1dclEiu7PR8QZ8cwU9rR4Zo/u4c1fmzY/7JURcf+Wsnq7v9sbEavLstz3eZ53bkT834i4ISIej4iXbi7NZ0fEZ8uy/G9TuC4AAAAAALpIURRnRsRvIuK1EfHRiPh/EfE/yrJcOeXnfqGi+1lB+iLiT8uyPOlZx7ef6H5ZWZa/3vz9uyPi4rIsjy6KYr+IeHLzpPeBEbEwIt5VluXioij+T0TcsN3NKO8qy/KbU704AAAAAADqb6J7dE/GBUVRvDOemQBfHRFnbj4+MyL+V1EUo/HMyPulZVku3nzu4oj4flEUl0TELyLiqp2QCwAAAACAGprURDcAAAAAAHSbnhd+CAAAAAAAdC9FNwAAAAAAWVN0AwAAAACQNUU3AAAAAABZU3QDAAAAAJA1RTcAAAAAAFlTdAMAAAAAkDVFNwAAAAAAWfv/AHjSkJTfg6UAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 1800x720 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 514
        },
        "id": "WyXpKiXJ_Aen",
        "outputId": "22d91bd6-a9d2-4ee6-c615-884bddf17801"
      },
      "source": [
        "msno.bar(train_df, labels=True,fontsize=10)"
      ],
      "execution_count": 275,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f95dbb6c5d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 275
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1728x720 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 990
        },
        "id": "__8dAuz3_K_y",
        "outputId": "da6cfca0-8154-4e52-e586-fd893de2739a"
      },
      "source": [
        "plt.figure(figsize=(15,15))\n",
        "sns.heatmap(train_df.isnull())"
      ],
      "execution_count": 276,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f95e32e84d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 276
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "B8ARMh1P_WxI",
        "outputId": "61c56b32-e174-4c12-baa8-07d0ca23e6cc"
      },
      "source": [
        "#finding out cols with significant no. of Missing Vals\n",
        "null_col_names = []\n",
        "\n",
        "for col in train_df.columns.values: # For each column\n",
        "  if train_df[col].isnull().sum()>20:\n",
        "   print('{col_name} has {num} null entries.'.format(col_name=col, num = str(train_df[col].isnull().sum())))\n",
        "   null_col_names.append(col)\n",
        "print('\\nColumns to delete: {}'.format(null_col_names))\n"
      ],
      "execution_count": 277,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "funder has 2793 null entries.\n",
            "installer has 2807 null entries.\n",
            "subvillage has 287 null entries.\n",
            "public_meeting has 2491 null entries.\n",
            "scheme_management has 2832 null entries.\n",
            "scheme_name has 21110 null entries.\n",
            "permit has 2336 null entries.\n",
            "\n",
            "Columns to delete: ['funder', 'installer', 'subvillage', 'public_meeting', 'scheme_management', 'scheme_name', 'permit']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "v6zfK1ExBitG",
        "outputId": "5f97d67f-a984-427f-c431-d5d6f502489c"
      },
      "source": [
        "#Finding out rows with significant no. of missing values ,we are putting a threshold of 10 missing values\n",
        "null_sum = train_df.isnull().sum(axis=1).tolist()\n",
        "null_row_indices = [index for index, val in enumerate(null_sum) if val > 10] \n",
        "for row in null_row_indices:\n",
        " print('Row {row} has  {num} null values'.format(row = row, num = str(null_sum[row]) ))\n",
        "print('\\n')\n",
        "print('Rows to delete: {}'.format(null_row_indices))"
      ],
      "execution_count": 278,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "\n",
            "Rows to delete: []\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "id": "jZxSqsC0E5nW",
        "outputId": "fbec88ba-3ecc-4b6c-cc85-2a978e0efb2e"
      },
      "source": [
        "#Checking for duplicates\n",
        "train_df_dupe = train_df[train_df.duplicated()]\n",
        "train_df_dupe"
      ],
      "execution_count": 279,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>date_recorded</th>\n",
              "      <th>funder</th>\n",
              "      <th>gps_height</th>\n",
              "      <th>installer</th>\n",
              "      <th>longitude</th>\n",
              "      <th>latitude</th>\n",
              "      <th>wpt_name</th>\n",
              "      <th>num_private</th>\n",
              "      <th>basin</th>\n",
              "      <th>subvillage</th>\n",
              "      <th>region</th>\n",
              "      <th>region_code</th>\n",
              "      <th>district_code</th>\n",
              "      <th>lga</th>\n",
              "      <th>ward</th>\n",
              "      <th>population</th>\n",
              "      <th>public_meeting</th>\n",
              "      <th>recorded_by</th>\n",
              "      <th>scheme_management</th>\n",
              "      <th>scheme_name</th>\n",
              "      <th>permit</th>\n",
              "      <th>construction_year</th>\n",
              "      <th>extraction_type</th>\n",
              "      <th>extraction_type_group</th>\n",
              "      <th>extraction_type_class</th>\n",
              "      <th>management</th>\n",
              "      <th>management_group</th>\n",
              "      <th>payment</th>\n",
              "      <th>payment_type</th>\n",
              "      <th>water_quality</th>\n",
              "      <th>quality_group</th>\n",
              "      <th>quantity</th>\n",
              "      <th>quantity_group</th>\n",
              "      <th>source</th>\n",
              "      <th>source_type</th>\n",
              "      <th>source_class</th>\n",
              "      <th>waterpoint_type</th>\n",
              "      <th>waterpoint_type_group</th>\n",
              "      <th>status_group</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "Empty DataFrame\n",
              "Columns: [id, amount_tsh, date_recorded, funder, gps_height, installer, longitude, latitude, wpt_name, num_private, basin, subvillage, region, region_code, district_code, lga, ward, population, public_meeting, recorded_by, scheme_management, scheme_name, permit, construction_year, extraction_type, extraction_type_group, extraction_type_class, management, management_group, payment, payment_type, water_quality, quality_group, quantity, quantity_group, source, source_type, source_class, waterpoint_type, waterpoint_type_group, status_group]\n",
              "Index: []"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 279
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "id": "pSuYRgokFFRX",
        "outputId": "f1259e7f-8999-446d-da5d-cd1b46e026af"
      },
      "source": [
        "train_df.status_group.value_counts().plot(kind='barh')"
      ],
      "execution_count": 280,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f95db9e6e10>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 280
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Qhkl323HKnZ4"
      },
      "source": [
        "train_df.construction_year = pd.to_numeric(train_df.construction_year)"
      ],
      "execution_count": 281,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9Kiw2DAMNe--"
      },
      "source": [
        "train_df.loc[train_df.construction_year <= 0, train_df.columns=='construction_year'] = 1950"
      ],
      "execution_count": 282,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o7Jy2GBpNyAJ"
      },
      "source": [
        "hist1=train_df[train_df.status_group == 'functional'].construction_year\n",
        "hist2=train_df[train_df.status_group == 'functional needs repair'].construction_year\n",
        "hist3=train_df[train_df.status_group == 'non functional'].construction_year"
      ],
      "execution_count": 283,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 362
        },
        "id": "q55scS9qOUb4",
        "outputId": "b90d6ef1-f020-4e1c-867a-e0a7d22f7b8f"
      },
      "source": [
        "n,b,p=plt.hist([hist1, hist2, hist3], stacked=True,range=[1950,2010])\n",
        "plt.legend(['functional','functional needs repair','non functional'],loc=0)\n",
        "plt.text(1952, 6000,'NO DATA',fontsize=17,rotation=90,color='black')\n",
        "plt.xlabel('Construction Year', fontsize=18)\n",
        "\n"
      ],
      "execution_count": 284,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/numpy/core/_asarray.py:83: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "  return array(a, dtype, copy=False, order=order)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'Construction Year')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 284
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "niBdlsoAbB2x"
      },
      "source": [
        "## Data Preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I_WHL-l189nu",
        "outputId": "c18c534d-0a5a-4dfa-aa84-c5f4394c31a4"
      },
      "source": [
        "train_df.funder.value_counts()"
      ],
      "execution_count": 285,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Government Of Tanzania    6780\n",
              "Danida                    2362\n",
              "Hesawa                    1639\n",
              "Rwssp                     1032\n",
              "Kkkt                       964\n",
              "                          ... \n",
              "Nyangere                     1\n",
              "Vwt                          1\n",
              "Mbwana Omari                 1\n",
              "Un/wfp                       1\n",
              "Rajab Seleman                1\n",
              "Name: funder, Length: 1651, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 285
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WQ5Z6TcKIMeM"
      },
      "source": [
        "def funder_wrangler(row):\n",
        "  '''Keep top 5 funders and setting rest as others'''\n",
        "  if row['funder']=='Government Of Tanzania':\n",
        "      return 'gov'\n",
        "  elif row['funder']=='Danida':\n",
        "      return 'danida'\n",
        "  elif row['funder']=='Hesawa':\n",
        "      return 'hesawa'\n",
        "  elif row['funder']=='Rwssp':\n",
        "      return 'rwssp'\n",
        "  elif row['funder']=='World Bank':\n",
        "      return 'world_bank'    \n",
        "  else:\n",
        "      return 'other'\n",
        "\n",
        "train_df['funder'] = train_df.apply(lambda row: funder_wrangler(row), axis=1)"
      ],
      "execution_count": 286,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OkA7h-QOLOJF"
      },
      "source": [
        "# Add column named 'status_group_vals' to allow the use of a pivot table to check differences\n",
        "# between the different funders.\n",
        "\n",
        "vals_to_replace = {'functional':2, 'functional needs repair':1,'non functional':0}\n",
        "\n",
        "train_df['status_group_vals']  = train_df.status_group.replace(vals_to_replace)"
      ],
      "execution_count": 287,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 808
        },
        "id": "3S55TlOmbLV7",
        "outputId": "a60249e3-664d-48ee-baf0-88095f07b614"
      },
      "source": [
        "train_df"
      ],
      "execution_count": 288,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>date_recorded</th>\n",
              "      <th>funder</th>\n",
              "      <th>gps_height</th>\n",
              "      <th>installer</th>\n",
              "      <th>longitude</th>\n",
              "      <th>latitude</th>\n",
              "      <th>wpt_name</th>\n",
              "      <th>num_private</th>\n",
              "      <th>basin</th>\n",
              "      <th>subvillage</th>\n",
              "      <th>region</th>\n",
              "      <th>region_code</th>\n",
              "      <th>district_code</th>\n",
              "      <th>lga</th>\n",
              "      <th>ward</th>\n",
              "      <th>population</th>\n",
              "      <th>public_meeting</th>\n",
              "      <th>recorded_by</th>\n",
              "      <th>scheme_management</th>\n",
              "      <th>scheme_name</th>\n",
              "      <th>permit</th>\n",
              "      <th>construction_year</th>\n",
              "      <th>extraction_type</th>\n",
              "      <th>extraction_type_group</th>\n",
              "      <th>extraction_type_class</th>\n",
              "      <th>management</th>\n",
              "      <th>management_group</th>\n",
              "      <th>payment</th>\n",
              "      <th>payment_type</th>\n",
              "      <th>water_quality</th>\n",
              "      <th>quality_group</th>\n",
              "      <th>quantity</th>\n",
              "      <th>quantity_group</th>\n",
              "      <th>source</th>\n",
              "      <th>source_type</th>\n",
              "      <th>source_class</th>\n",
              "      <th>waterpoint_type</th>\n",
              "      <th>waterpoint_type_group</th>\n",
              "      <th>status_group</th>\n",
              "      <th>status_group_vals</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>7411</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2013-03-20</td>\n",
              "      <td>other</td>\n",
              "      <td>1203</td>\n",
              "      <td>CES</td>\n",
              "      <td>37.228664</td>\n",
              "      <td>-3.256505</td>\n",
              "      <td>Kwa Abdalla</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>Mafurini</td>\n",
              "      <td>Kilimanjaro</td>\n",
              "      <td>3</td>\n",
              "      <td>5</td>\n",
              "      <td>Hai</td>\n",
              "      <td>Machame Kaskazini</td>\n",
              "      <td>90</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>Water Board</td>\n",
              "      <td>Machame water supply</td>\n",
              "      <td>True</td>\n",
              "      <td>2008</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>water board</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay per bucket</td>\n",
              "      <td>per bucket</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>enough</td>\n",
              "      <td>enough</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>18014</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2011-02-21</td>\n",
              "      <td>other</td>\n",
              "      <td>1409</td>\n",
              "      <td>CDTF</td>\n",
              "      <td>34.615481</td>\n",
              "      <td>-9.972601</td>\n",
              "      <td>none</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Nyasa</td>\n",
              "      <td>Ituni</td>\n",
              "      <td>Iringa</td>\n",
              "      <td>11</td>\n",
              "      <td>5</td>\n",
              "      <td>Ludewa</td>\n",
              "      <td>Luana</td>\n",
              "      <td>70</td>\n",
              "      <td>False</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>Private operator</td>\n",
              "      <td>Mgowel</td>\n",
              "      <td>False</td>\n",
              "      <td>1984</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>private operator</td>\n",
              "      <td>commercial</td>\n",
              "      <td>never pay</td>\n",
              "      <td>never pay</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>enough</td>\n",
              "      <td>enough</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>26103</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2013-01-04</td>\n",
              "      <td>gov</td>\n",
              "      <td>1572</td>\n",
              "      <td>DWE</td>\n",
              "      <td>36.658778</td>\n",
              "      <td>-3.321529</td>\n",
              "      <td>Kwa Loishiye</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>Melebuki</td>\n",
              "      <td>Arusha</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>Arusha Rural</td>\n",
              "      <td>Kimnyaki</td>\n",
              "      <td>150</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>Saitero olosaita pipe line</td>\n",
              "      <td>True</td>\n",
              "      <td>2002</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>vwc</td>\n",
              "      <td>user-group</td>\n",
              "      <td>never pay</td>\n",
              "      <td>never pay</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>non functional</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>31140</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2011-04-03</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>District Council</td>\n",
              "      <td>33.628025</td>\n",
              "      <td>-8.860795</td>\n",
              "      <td>Mwanzechele</td>\n",
              "      <td>0</td>\n",
              "      <td>Rufiji</td>\n",
              "      <td>Darajani</td>\n",
              "      <td>Mbeya</td>\n",
              "      <td>12</td>\n",
              "      <td>2</td>\n",
              "      <td>Mbeya Rural</td>\n",
              "      <td>Inyala</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>NaN</td>\n",
              "      <td>True</td>\n",
              "      <td>1950</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>wug</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay annually</td>\n",
              "      <td>annually</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>river</td>\n",
              "      <td>river/lake</td>\n",
              "      <td>surface</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>12286</td>\n",
              "      <td>300.0</td>\n",
              "      <td>2013-03-08</td>\n",
              "      <td>other</td>\n",
              "      <td>1283</td>\n",
              "      <td>District Council</td>\n",
              "      <td>37.655509</td>\n",
              "      <td>-3.669777</td>\n",
              "      <td>Kwa Kidaya</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>Mwira A</td>\n",
              "      <td>Kilimanjaro</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>Mwanga</td>\n",
              "      <td>Shinghatini</td>\n",
              "      <td>34</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>Churu water supply</td>\n",
              "      <td>False</td>\n",
              "      <td>2007</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>vwc</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay monthly</td>\n",
              "      <td>monthly</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44545</th>\n",
              "      <td>43878</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2013-01-23</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>TAWASA</td>\n",
              "      <td>32.383715</td>\n",
              "      <td>-3.457624</td>\n",
              "      <td>Imalabuluhi</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Tanganyika</td>\n",
              "      <td>Bukenge</td>\n",
              "      <td>Shinyanga</td>\n",
              "      <td>17</td>\n",
              "      <td>3</td>\n",
              "      <td>Kahama</td>\n",
              "      <td>Segese</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>True</td>\n",
              "      <td>1950</td>\n",
              "      <td>nira/tanira</td>\n",
              "      <td>nira/tanira</td>\n",
              "      <td>handpump</td>\n",
              "      <td>wug</td>\n",
              "      <td>user-group</td>\n",
              "      <td>unknown</td>\n",
              "      <td>unknown</td>\n",
              "      <td>milky</td>\n",
              "      <td>milky</td>\n",
              "      <td>enough</td>\n",
              "      <td>enough</td>\n",
              "      <td>shallow well</td>\n",
              "      <td>shallow well</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44546</th>\n",
              "      <td>50718</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2011-07-25</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>DED</td>\n",
              "      <td>31.615509</td>\n",
              "      <td>-2.738682</td>\n",
              "      <td>Ilyamchele Center</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Victoria</td>\n",
              "      <td>Ilyamchele Center</td>\n",
              "      <td>Kagera</td>\n",
              "      <td>18</td>\n",
              "      <td>8</td>\n",
              "      <td>Chato</td>\n",
              "      <td>Katende</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>WUA</td>\n",
              "      <td>NaN</td>\n",
              "      <td>True</td>\n",
              "      <td>1950</td>\n",
              "      <td>afridev</td>\n",
              "      <td>afridev</td>\n",
              "      <td>handpump</td>\n",
              "      <td>wua</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay monthly</td>\n",
              "      <td>monthly</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>enough</td>\n",
              "      <td>enough</td>\n",
              "      <td>hand dtw</td>\n",
              "      <td>borehole</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>non functional</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44547</th>\n",
              "      <td>67177</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2013-02-23</td>\n",
              "      <td>gov</td>\n",
              "      <td>969</td>\n",
              "      <td>Government</td>\n",
              "      <td>35.134412</td>\n",
              "      <td>-10.676062</td>\n",
              "      <td>Kwa Mzee Maondo</td>\n",
              "      <td>0</td>\n",
              "      <td>Ruvuma / Southern Coast</td>\n",
              "      <td>Masango</td>\n",
              "      <td>Ruvuma</td>\n",
              "      <td>10</td>\n",
              "      <td>3</td>\n",
              "      <td>Mbinga</td>\n",
              "      <td>Kigonsera</td>\n",
              "      <td>60</td>\n",
              "      <td>NaN</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>Government</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1977</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>vwc</td>\n",
              "      <td>user-group</td>\n",
              "      <td>unknown</td>\n",
              "      <td>unknown</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>seasonal</td>\n",
              "      <td>seasonal</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44548</th>\n",
              "      <td>2744</td>\n",
              "      <td>200.0</td>\n",
              "      <td>2013-03-26</td>\n",
              "      <td>other</td>\n",
              "      <td>1269</td>\n",
              "      <td>ADRA/Government</td>\n",
              "      <td>37.886287</td>\n",
              "      <td>-4.390882</td>\n",
              "      <td>Miembeni</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>Mng'Ende</td>\n",
              "      <td>Kilimanjaro</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>Same</td>\n",
              "      <td>Suji</td>\n",
              "      <td>80</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>Water authority</td>\n",
              "      <td>Kiumba water supply</td>\n",
              "      <td>True</td>\n",
              "      <td>2011</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>gravity</td>\n",
              "      <td>vwc</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay monthly</td>\n",
              "      <td>monthly</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44549</th>\n",
              "      <td>10262</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2013-01-26</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Consulting Engineer</td>\n",
              "      <td>32.976560</td>\n",
              "      <td>-4.262149</td>\n",
              "      <td>Itunda A Mkwajuni</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Tanganyika</td>\n",
              "      <td>Itunda A</td>\n",
              "      <td>Tabora</td>\n",
              "      <td>14</td>\n",
              "      <td>1</td>\n",
              "      <td>Nzega</td>\n",
              "      <td>Karitu</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>GeoData Consultants Ltd</td>\n",
              "      <td>VWC</td>\n",
              "      <td>NaN</td>\n",
              "      <td>False</td>\n",
              "      <td>1950</td>\n",
              "      <td>nira/tanira</td>\n",
              "      <td>nira/tanira</td>\n",
              "      <td>handpump</td>\n",
              "      <td>vwc</td>\n",
              "      <td>user-group</td>\n",
              "      <td>pay when scheme fails</td>\n",
              "      <td>on failure</td>\n",
              "      <td>soft</td>\n",
              "      <td>good</td>\n",
              "      <td>enough</td>\n",
              "      <td>enough</td>\n",
              "      <td>shallow well</td>\n",
              "      <td>shallow well</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>functional</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>44550 rows × 42 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "          id  amount_tsh  ...    status_group status_group_vals\n",
              "0       7411         7.0  ...      functional                 2\n",
              "1      18014         0.0  ...      functional                 2\n",
              "2      26103         0.0  ...  non functional                 0\n",
              "3      31140         0.0  ...      functional                 2\n",
              "4      12286       300.0  ...      functional                 2\n",
              "...      ...         ...  ...             ...               ...\n",
              "44545  43878         0.0  ...      functional                 2\n",
              "44546  50718         0.0  ...  non functional                 0\n",
              "44547  67177         0.0  ...      functional                 2\n",
              "44548   2744       200.0  ...      functional                 2\n",
              "44549  10262         0.0  ...      functional                 2\n",
              "\n",
              "[44550 rows x 42 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 288
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LJe5XlCHLdng",
        "outputId": "bd414fe3-7570-433f-86ef-1f6f427299e7"
      },
      "source": [
        "piv_table = pd.pivot_table(train_df,index=['funder','status_group'],values='status_group_vals', aggfunc='count')\n",
        "print(piv_table)"
      ],
      "execution_count": 289,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                                    status_group_vals\n",
            "funder     status_group                              \n",
            "danida     functional                            1282\n",
            "           functional needs repair                119\n",
            "           non functional                         961\n",
            "gov        functional                            2781\n",
            "           functional needs repair                519\n",
            "...                                               ...\n",
            "rwssp      functional needs repair                 80\n",
            "           non functional                         343\n",
            "world_bank functional                             399\n",
            "           functional needs repair                 73\n",
            "           non functional                         490\n",
            "\n",
            "[18 rows x 1 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1G0VYaroLh97",
        "outputId": "de287efd-6682-40be-f206-2371fb879c72"
      },
      "source": [
        "total_danida = piv_table.status_group_vals[0] + piv_table.status_group_vals[1] + piv_table.status_group_vals[2]\n",
        "percent_functional_danida = (piv_table.status_group_vals[0] / total_danida) * 100\n",
        "\n",
        "total_gov = piv_table.status_group_vals[3] + piv_table.status_group_vals[4] + piv_table.status_group_vals[5]\n",
        "percent_functional_gov = (piv_table.status_group_vals[3] / total_gov) * 100\n",
        "\n",
        "total_hesawa = piv_table.status_group_vals[6] + piv_table.status_group_vals[7] + piv_table.status_group_vals[8]\n",
        "percent_functional_hesawa = (piv_table.status_group_vals[6] / total_hesawa) * 100\n",
        "\n",
        "total_other = piv_table.status_group_vals[9] + piv_table.status_group_vals[10] + piv_table.status_group_vals[11]\n",
        "percent_functional_non_gov = (piv_table.status_group_vals[9] / total_other) * 100\n",
        "\n",
        "total_rwssp = piv_table.status_group_vals[12] + piv_table.status_group_vals[13] + piv_table.status_group_vals[14]\n",
        "percent_functional_rwssp = (piv_table.status_group_vals[12] / total_rwssp) * 100\n",
        "\n",
        "total_world_bank = piv_table.status_group_vals[15] + piv_table.status_group_vals[16] + piv_table.status_group_vals[17]\n",
        "percent_functional_world_bank = (piv_table.status_group_vals[15] / total_world_bank) * 100\n",
        "\n",
        "print('Percent functional danida: ', round(percent_functional_danida,3))\n",
        "print('Percent functional gov: ', round(percent_functional_gov,3))\n",
        "print('Percent functional hesawa: ', round(percent_functional_hesawa,3))\n",
        "print('Percent functional non gov: ', round(percent_functional_non_gov,3))\n",
        "print('Percent functional rwssp: ', round(percent_functional_rwssp,3))\n",
        "print('Percent functional world bank: ', round(percent_functional_world_bank,3))"
      ],
      "execution_count": 290,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Percent functional danida:  54.276\n",
            "Percent functional gov:  41.018\n",
            "Percent functional hesawa:  42.221\n",
            "Percent functional non gov:  58.005\n",
            "Percent functional rwssp:  59.012\n",
            "Percent functional world bank:  41.476\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "B4MQ0pG-MAR6",
        "outputId": "8408fcd4-2bf6-4762-ab07-5675fd7fcb95"
      },
      "source": [
        "# There are some clear differences here that will hopefully improve the model. The next feature\n",
        "# to inspect is 'installer'.\n",
        "\n",
        "train_df.installer.value_counts()"
      ],
      "execution_count": 291,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DWE              13005\n",
              "Government        1365\n",
              "RWE                876\n",
              "Commu              806\n",
              "DANIDA             788\n",
              "                 ...  \n",
              "MANGO TREE           1\n",
              "Raramataki           1\n",
              "Privat               1\n",
              "Goldmain             1\n",
              "Meru Concrete        1\n",
              "Name: installer, Length: 1854, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 291
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SX6aCI-mNXrh"
      },
      "source": [
        "# Create a function to reduce the amount of dummy columns needed whilst maintaining the \n",
        "# information contained in the column.\n",
        "\n",
        "def installer_wrangler(row):\n",
        "    '''Keep top 5 values and set the rest to 'other'''\n",
        "    if row['installer']=='DWE':\n",
        "        return 'dwe'\n",
        "    elif row['installer']=='Government':\n",
        "        return 'gov'\n",
        "    elif row['installer']=='RWE':\n",
        "        return 'rwe'\n",
        "    elif row['installer']=='Commu':\n",
        "        return 'commu'\n",
        "    elif row['installer']=='DANIDA':\n",
        "        return 'danida'\n",
        "    else:\n",
        "        return 'other'  \n",
        "\n",
        "train_df['installer'] = train_df.apply(lambda row: installer_wrangler(row), axis=1)"
      ],
      "execution_count": 292,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 455
        },
        "id": "NOlsr1fibvnw",
        "outputId": "8b023064-f5e7-458c-93e4-12844ebb77ac"
      },
      "source": [
        "piv_table = pd.pivot_table(train_df,index=['installer','status_group'],values='status_group_vals', aggfunc='count')\n",
        "piv_table"
      ],
      "execution_count": 293,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th>status_group_vals</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>installer</th>\n",
              "      <th>status_group</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">commu</th>\n",
              "      <th>functional</th>\n",
              "      <td>546</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>27</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>233</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"2\" valign=\"top\">danida</th>\n",
              "      <th>functional</th>\n",
              "      <td>404</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>64</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"2\" valign=\"top\">other</th>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>1654</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>10503</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">rwe</th>\n",
              "      <th>functional</th>\n",
              "      <td>229</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>99</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>548</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>18 rows × 1 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                   status_group_vals\n",
              "installer status_group                              \n",
              "commu     functional                             546\n",
              "          functional needs repair                 27\n",
              "          non functional                         233\n",
              "danida    functional                             404\n",
              "          functional needs repair                 64\n",
              "...                                              ...\n",
              "other     functional needs repair               1654\n",
              "          non functional                       10503\n",
              "rwe       functional                             229\n",
              "          functional needs repair                 99\n",
              "          non functional                         548\n",
              "\n",
              "[18 rows x 1 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 293
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Mc0Z1VAETFeO",
        "outputId": "b281f301-b7a2-43ea-9291-a46f3baea062"
      },
      "source": [
        "# As was the case with 'funder' there are some clear differences here that will hopefully\n",
        "# improve the model. The next feature to inspect is 'subvillage'.\n",
        "\n",
        "print(df.subvillage.value_counts())"
      ],
      "execution_count": 294,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Madukani     508\n",
            "Shuleni      506\n",
            "Majengo      502\n",
            "Kati         373\n",
            "Mtakuja      262\n",
            "            ... \n",
            "Mwamamumi      1\n",
            "Kiware         1\n",
            "Naamalasi      1\n",
            "Qatabela       1\n",
            "Nyunjwi        1\n",
            "Name: subvillage, Length: 19287, dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "w2EHSXxSTrTI",
        "outputId": "0fcc0b6a-d523-4330-e346-6574c7bac819"
      },
      "source": [
        "print('Number of villages: ', len(df.subvillage.value_counts()))"
      ],
      "execution_count": 295,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Number of villages:  19287\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KMueahr_T01R"
      },
      "source": [
        "# 19287 unique values implies that the largest values don't dominate as was the case with installer and\n",
        "# funder. It's probably not worth creating dummy variables for the top 5. I'll drop this one but\n",
        "# feel free to experiment here.\n",
        "if 'subvillage' in train_df.columns.values:\n",
        "  train_df = train_df.drop('subvillage', axis=1)"
      ],
      "execution_count": 296,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zhM_TwM8UI_2",
        "outputId": "083c996e-7e20-4984-9f6c-096297da852b"
      },
      "source": [
        "# Let's investigate the next column containg null data: 'public_meeting'.\n",
        "\n",
        "print(train_df.public_meeting.value_counts(),\"\\n\")\n",
        "# We only have two values here: true and false. This one can stay but we'll have to replace \n",
        "# the unknown data with a string value.\n",
        "\n",
        "train_df.public_meeting = train_df.public_meeting.fillna('Unknown')\n",
        "\n",
        "# 'scheme_management' is up next.\n",
        "\n",
        "train_df.scheme_management.value_counts()"
      ],
      "execution_count": 297,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "True     38212\n",
            "False     3847\n",
            "Name: public_meeting, dtype: int64 \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "VWC                 27651\n",
              "WUG                  3906\n",
              "Water authority      2361\n",
              "WUA                  2166\n",
              "Water Board          2078\n",
              "Parastatal           1272\n",
              "Private operator      792\n",
              "Company               790\n",
              "Other                 575\n",
              "SWC                    69\n",
              "Trust                  57\n",
              "None                    1\n",
              "Name: scheme_management, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 297
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9Ve2H7xedIf_"
      },
      "source": [
        "# Create a function to reduce the amount of dummy columns needed whilst maintaining the \n",
        "# information contained in the column.\n",
        "\n",
        "def scheme_wrangler(row):\n",
        "    '''Keep top 5 values and set the rest to 'other'. '''\n",
        "    if row['scheme_management']=='VWC':\n",
        "        return 'vwc'\n",
        "    elif row['scheme_management']=='WUG':\n",
        "        return 'wug'\n",
        "    elif row['scheme_management']=='Water authority':\n",
        "        return 'wtr_auth'\n",
        "    elif row['scheme_management']=='WUA':\n",
        "        return 'wua'\n",
        "    elif row['scheme_management']=='Water Board':\n",
        "        return 'wtr_brd'\n",
        "    else:\n",
        "        return 'other'\n",
        "\n",
        "train_df['scheme_management'] = train_df.apply(lambda row: scheme_wrangler(row), axis=1)"
      ],
      "execution_count": 303,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 645
        },
        "id": "qJ4RfqWfUYOC",
        "outputId": "32db95db-57f2-4375-ddf0-b66dbd56a852"
      },
      "source": [
        "pd.set_option('display.max_rows', None)\n",
        "piv_table = pd.pivot_table(train_df, index=['scheme_management', 'status_group'],values='status_group_vals', aggfunc='count')\n",
        "piv_table.to_excel(\"Schememanagement.xlsx\")\n",
        "display(piv_table)\n",
        "pd.set_option('display.max_rows', 15)"
      ],
      "execution_count": 304,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th>status_group_vals</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>scheme_management</th>\n",
              "      <th>status_group</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">other</th>\n",
              "      <th>functional</th>\n",
              "      <td>3455</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>377</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>2556</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">vwc</th>\n",
              "      <th>functional</th>\n",
              "      <td>14242</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>1777</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>11632</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">wtr_auth</th>\n",
              "      <th>functional</th>\n",
              "      <td>1208</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>325</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>828</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">wtr_brd</th>\n",
              "      <th>functional</th>\n",
              "      <td>1555</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>75</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>448</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">wua</th>\n",
              "      <th>functional</th>\n",
              "      <td>1494</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>183</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>489</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">wug</th>\n",
              "      <th>functional</th>\n",
              "      <td>2240</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>501</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>1165</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                           status_group_vals\n",
              "scheme_management status_group                              \n",
              "other             functional                            3455\n",
              "                  functional needs repair                377\n",
              "                  non functional                        2556\n",
              "vwc               functional                           14242\n",
              "                  functional needs repair               1777\n",
              "                  non functional                       11632\n",
              "wtr_auth          functional                            1208\n",
              "                  functional needs repair                325\n",
              "                  non functional                         828\n",
              "wtr_brd           functional                            1555\n",
              "                  functional needs repair                 75\n",
              "                  non functional                         448\n",
              "wua               functional                            1494\n",
              "                  functional needs repair                183\n",
              "                  non functional                         489\n",
              "wug               functional                            2240\n",
              "                  functional needs repair                501\n",
              "                  non functional                        1165"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7KjhywZGasJ0",
        "outputId": "188aee7f-2e62-4c7f-be8e-e8ac52798367"
      },
      "source": [
        "total_other = piv_table.status_group_vals[0] + piv_table.status_group_vals[1] + piv_table.status_group_vals[2]\n",
        "percent_functional_other = (piv_table.status_group_vals[0] / total_other) * 100\n",
        "\n",
        "total_vwc = piv_table.status_group_vals[3] + piv_table.status_group_vals[4] + piv_table.status_group_vals[5]\n",
        "percent_functional_vwc = (piv_table.status_group_vals[3] / total_vwc) * 100\n",
        "\n",
        "total_wtr_auth = piv_table.status_group_vals[6] + piv_table.status_group_vals[7] + piv_table.status_group_vals[8]\n",
        "percent_functional_wtr_auth = (piv_table.status_group_vals[6] / total_wtr_auth) * 100\n",
        "\n",
        "total_wtr_brd = piv_table.status_group_vals[9] + piv_table.status_group_vals[10] + piv_table.status_group_vals[11]\n",
        "percent_functional_wtr_brd = (piv_table.status_group_vals[9] / total_wtr_brd) * 100\n",
        "\n",
        "total_wua = piv_table.status_group_vals[12] + piv_table.status_group_vals[13] + piv_table.status_group_vals[14]\n",
        "percent_functional_wua = (piv_table.status_group_vals[12] / total_wua) * 100\n",
        "\n",
        "total_wug = piv_table.status_group_vals[15] + piv_table.status_group_vals[16] + piv_table.status_group_vals[17]\n",
        "percent_functional_wug = (piv_table.status_group_vals[15] / total_wug) * 100\n",
        "\n",
        "print('Percent functional other: ', round(percent_functional_other,3))\n",
        "print('Percent functional vwc: ', round(percent_functional_vwc,3))\n",
        "print('Percent functional water authority: ', round(percent_functional_wtr_auth,3))\n",
        "print('Percent functional water board: ', round(percent_functional_wtr_brd,3))\n",
        "print('Percent functional wua: ', round(percent_functional_wua,3))\n",
        "print('Percent functional wug: ', round(percent_functional_wug,3))"
      ],
      "execution_count": 305,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Percent functional other:  54.086\n",
            "Percent functional vwc:  51.506\n",
            "Percent functional water authority:  51.165\n",
            "Percent functional water board:  74.832\n",
            "Percent functional wua:  68.975\n",
            "Percent functional wug:  57.348\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oR6rbYoohBqi",
        "outputId": "938a2fbc-a002-47ee-d3b8-1ce70b2df400"
      },
      "source": [
        "train_df.scheme_name.value_counts()"
      ],
      "execution_count": 306,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "K                                      528\n",
              "None                                   494\n",
              "Borehole                               402\n",
              "M                                      310\n",
              "Chalinze wate                          300\n",
              "                                      ... \n",
              "Gwarama point  sources water supply      1\n",
              "Mtumbei mpopera                          1\n",
              "Marua mshatie water supply               1\n",
              "Mand                                     1\n",
              "Mongwe r                                 1\n",
              "Name: scheme_name, Length: 2506, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 306
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eBfX5JA2hFLH"
      },
      "source": [
        "len(train_df.scheme_name.unique())\n",
        "\n",
        "# Lots of factors and the top 5 or so only represent a fraction of the total values. Probably \n",
        "# safe to drop this column.\n",
        "\n",
        "train_df = train_df.drop('scheme_name', axis=1)"
      ],
      "execution_count": 307,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cSxkgoqEhRDY",
        "outputId": "684bf2a0-c2af-4ca5-dfcb-066f3e8a0bc2"
      },
      "source": [
        "# The final column containing nulls is 'permit'.\n",
        "\n",
        "train_df.permit.value_counts()"
      ],
      "execution_count": 308,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True     29057\n",
              "False    13157\n",
              "Name: permit, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 308
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zSr0d0wShXkQ"
      },
      "source": [
        "# We only have two values here: true and false. This one can stay but we'll have to replace \n",
        "# the unknown data with a string value.\n",
        "\n",
        "train_df.permit = train_df.permit.fillna('Unknown')"
      ],
      "execution_count": 309,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4BRlkznLhkjh",
        "outputId": "3ae49cad-5c8f-46fd-d83f-e031203aaa89"
      },
      "source": [
        "pd.set_option('display.max_rows', None)\n",
        "print(train_df.apply(lambda x: sum(x.isnull())))\n",
        "pd.set_option('display.max_rows', 15)"
      ],
      "execution_count": 312,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "id                       0\n",
            "amount_tsh               0\n",
            "date_recorded            0\n",
            "funder                   0\n",
            "gps_height               0\n",
            "installer                0\n",
            "longitude                0\n",
            "latitude                 0\n",
            "wpt_name                 0\n",
            "num_private              0\n",
            "basin                    0\n",
            "region                   0\n",
            "region_code              0\n",
            "district_code            0\n",
            "lga                      0\n",
            "ward                     0\n",
            "population               0\n",
            "public_meeting           0\n",
            "recorded_by              0\n",
            "scheme_management        0\n",
            "permit                   0\n",
            "construction_year        0\n",
            "extraction_type          0\n",
            "extraction_type_group    0\n",
            "extraction_type_class    0\n",
            "management               0\n",
            "management_group         0\n",
            "payment                  0\n",
            "payment_type             0\n",
            "water_quality            0\n",
            "quality_group            0\n",
            "quantity                 0\n",
            "quantity_group           0\n",
            "source                   0\n",
            "source_type              0\n",
            "source_class             0\n",
            "waterpoint_type          0\n",
            "waterpoint_type_group    0\n",
            "status_group             0\n",
            "status_group_vals        0\n",
            "dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OLb5yx9phs8C",
        "outputId": "c8b7d0ad-bb5e-4175-baa2-7158d3014d46"
      },
      "source": [
        "# Excellent! Now there are no nulls in the data set. We can move on to look at columns with \n",
        "# string values and modify or remove them as we see fit.\n",
        "\n",
        "str_cols = train_df.select_dtypes(include = ['object'])\n",
        "str_cols.apply(lambda x: len(x.unique()))"
      ],
      "execution_count": 313,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "date_recorded              346\n",
              "funder                       6\n",
              "installer                    6\n",
              "wpt_name                 28991\n",
              "basin                        9\n",
              "                         ...  \n",
              "source_type                  7\n",
              "source_class                 3\n",
              "waterpoint_type              7\n",
              "waterpoint_type_group        6\n",
              "status_group                 3\n",
              "Length: 29, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 313
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vPZ-AIMIiDAC",
        "outputId": "41bf829f-edad-4d94-8ae3-b4a4caf437d2"
      },
      "source": [
        "# 'Date recorded'\n",
        "\n",
        "train_df.date_recorded.describe()"
      ],
      "execution_count": 314,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count          44550\n",
              "unique           346\n",
              "top       2011-03-15\n",
              "freq             423\n",
              "Name: date_recorded, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 314
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6zjCyLX1iN99",
        "outputId": "f82d63d5-2a9c-47f9-f757-620af1ca3f0c"
      },
      "source": [
        "# Let's first convert the column to type datetime. Then change the column to represent the \n",
        "# number of days since the most recently recorded datum we have. The idea being that more \n",
        "# recently recorded pumps might be more likely to be functional than non-functional. \n",
        "\n",
        "train_df.date_recorded = pd.to_datetime(train_df.date_recorded)\n",
        "train_df.date_recorded.describe()"
      ],
      "execution_count": 315,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:6: FutureWarning: Treating datetime data as categorical rather than numeric in `.describe` is deprecated and will be removed in a future version of pandas. Specify `datetime_is_numeric=True` to silence this warning and adopt the future behavior now.\n",
            "  \n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count                   44550\n",
              "unique                    346\n",
              "top       2011-03-15 00:00:00\n",
              "freq                      423\n",
              "first     2002-10-14 00:00:00\n",
              "last      2013-12-03 00:00:00\n",
              "Name: date_recorded, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 315
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ad4KikVciYjp",
        "outputId": "ac7a2820-a66b-4a4a-b6d3-505f3c14f325"
      },
      "source": [
        "# The most recent data is 2013-12-03. Subtract each date from this point to obtain a \n",
        "# 'days_since_recorded' column.\n",
        "\n",
        "train_df.date_recorded = pd.datetime(2013, 12, 3) - pd.to_datetime(train_df.date_recorded)\n",
        "train_df.columns = ['days_since_recorded' if x=='date_recorded' else x for x in train_df.columns]\n",
        "train_df.days_since_recorded = train_df.days_since_recorded.astype('timedelta64[D]').astype(int)\n",
        "train_df.days_since_recorded.describe()"
      ],
      "execution_count": 316,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:4: FutureWarning: The pandas.datetime class is deprecated and will be removed from pandas in a future version. Import from datetime module instead.\n",
            "  after removing the cwd from sys.path.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count    44550.000000\n",
              "mean       615.289809\n",
              "std        333.898570\n",
              "min          0.000000\n",
              "25%        297.000000\n",
              "50%        420.000000\n",
              "75%        977.000000\n",
              "max       4068.000000\n",
              "Name: days_since_recorded, dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 316
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AeREsl2FirwD",
        "outputId": "6887c671-55ad-431f-c3a1-d36918938a6d"
      },
      "source": [
        "# There's a wide range of data here hopefully it will help improve the predictive power of our\n",
        "# models. Next up for inspection is 'wpt_name' (Name of the waterpoint if there is one).\n",
        "\n",
        "df.wpt_name.value_counts()\n",
        "\n"
      ],
      "execution_count": 317,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "none                     3563\n",
              "Shuleni                  1748\n",
              "Zahanati                  830\n",
              "Msikitini                 535\n",
              "Kanisani                  323\n",
              "                         ... \n",
              "Kwa Martini Moheda          1\n",
              "Rwatankororo                1\n",
              "Dulu                        1\n",
              "Kwa Mkwenya                 1\n",
              "Kwa Jerry Mwambambale       1\n",
              "Name: wpt_name, Length: 37400, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 317
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D_i6jtGqroLd"
      },
      "source": [
        "# Due to the huge number of factors and the lack of a clear dominating value I'll drop this.\n",
        "# I may come back and include the top 5 later. Next up is 'basin'.\n",
        "\n",
        "train_df = train_df.drop('wpt_name', axis=1)"
      ],
      "execution_count": 318,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YNgUfgJTtJM2",
        "outputId": "4a82569c-ef4a-4ef6-872d-1de1356e5338"
      },
      "source": [
        "train_df.basin.value_counts()"
      ],
      "execution_count": 319,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Lake Victoria              7623\n",
              "Pangani                    6703\n",
              "Rufiji                     6045\n",
              "Internal                   5855\n",
              "Lake Tanganyika            4775\n",
              "Wami / Ruvu                4430\n",
              "Lake Nyasa                 3902\n",
              "Ruvuma / Southern Coast    3381\n",
              "Lake Rukwa                 1836\n",
              "Name: basin, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 319
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 927
        },
        "id": "CcbJgtS7tN5s",
        "outputId": "b1cdb809-7b1c-4394-e5ad-4949e313fea8"
      },
      "source": [
        "pd.set_option('display.max_rows', None)\n",
        "piv_table = pd.pivot_table(train_df, index=['basin', 'status_group'],values=['status_group_vals'], aggfunc='count')\n",
        "display(piv_table)\n",
        "pd.set_option('display.max_rows', 15)"
      ],
      "execution_count": 322,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th>status_group_vals</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>basin</th>\n",
              "      <th>status_group</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Internal</th>\n",
              "      <th>functional</th>\n",
              "      <td>3379</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>420</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>2056</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Lake Nyasa</th>\n",
              "      <th>functional</th>\n",
              "      <td>2525</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>195</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>1182</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Lake Rukwa</th>\n",
              "      <th>functional</th>\n",
              "      <td>748</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>203</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>885</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Lake Tanganyika</th>\n",
              "      <th>functional</th>\n",
              "      <td>2318</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>545</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>1912</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Lake Victoria</th>\n",
              "      <th>functional</th>\n",
              "      <td>3793</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>741</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>3089</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Pangani</th>\n",
              "      <th>functional</th>\n",
              "      <td>4021</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>352</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>2330</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Rufiji</th>\n",
              "      <th>functional</th>\n",
              "      <td>3836</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>338</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>1871</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Ruvuma / Southern Coast</th>\n",
              "      <th>functional</th>\n",
              "      <td>1268</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>237</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>1876</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">Wami / Ruvu</th>\n",
              "      <th>functional</th>\n",
              "      <td>2306</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>functional needs repair</th>\n",
              "      <td>207</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>non functional</th>\n",
              "      <td>1917</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                 status_group_vals\n",
              "basin                   status_group                              \n",
              "Internal                functional                            3379\n",
              "                        functional needs repair                420\n",
              "                        non functional                        2056\n",
              "Lake Nyasa              functional                            2525\n",
              "                        functional needs repair                195\n",
              "                        non functional                        1182\n",
              "Lake Rukwa              functional                             748\n",
              "                        functional needs repair                203\n",
              "                        non functional                         885\n",
              "Lake Tanganyika         functional                            2318\n",
              "                        functional needs repair                545\n",
              "                        non functional                        1912\n",
              "Lake Victoria           functional                            3793\n",
              "                        functional needs repair                741\n",
              "                        non functional                        3089\n",
              "Pangani                 functional                            4021\n",
              "                        functional needs repair                352\n",
              "                        non functional                        2330\n",
              "Rufiji                  functional                            3836\n",
              "                        functional needs repair                338\n",
              "                        non functional                        1871\n",
              "Ruvuma / Southern Coast functional                            1268\n",
              "                        functional needs repair                237\n",
              "                        non functional                        1876\n",
              "Wami / Ruvu             functional                            2306\n",
              "                        functional needs repair                207\n",
              "                        non functional                        1917"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IAjNZSxEtxS6"
      },
      "source": [
        "\n",
        "\n",
        "Most basins have have more functional than non-functional pumps. Lake Rukwa and Ruvuma don't. All the values are over 2000 so this looks like a good feature to keep. Region will be considered next.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JtWhlzIjtVhc",
        "outputId": "07de19e5-7887-4647-c8e6-759225f8e7a7"
      },
      "source": [
        "train_df.region.value_counts()"
      ],
      "execution_count": 323,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Iringa           4017\n",
              "Shinyanga        3719\n",
              "Mbeya            3609\n",
              "Kilimanjaro      3307\n",
              "Morogoro         2961\n",
              "                 ... \n",
              "Rukwa            1353\n",
              "Mtwara           1308\n",
              "Manyara          1172\n",
              "Lindi            1152\n",
              "Dar es Salaam     593\n",
              "Name: region, Length: 21, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 323
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FlePbziGt4fD"
      },
      "source": [
        "# Both basin, lga, ward and region contain geographical information so there is a risk of them being\n",
        "# highly correlated with each other. I'll drop then for now.\n",
        "# They could be be worth including though, so I may come back to them.\n",
        "train_df = train_df.drop(['region', 'lga', 'ward'], axis=1)"
      ],
      "execution_count": 324,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pVrWdfF4uDJB",
        "outputId": "5152f445-f19d-4b5c-c2a5-f65ddbfd223c"
      },
      "source": [
        "train_df.recorded_by.value_counts()"
      ],
      "execution_count": 325,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GeoData Consultants Ltd    44550\n",
              "Name: recorded_by, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 325
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z7IJ6Lq9uHq5"
      },
      "source": [
        "# All data points have the same value so this offers no information that would help build our\n",
        "# model.\n",
        "train_df = train_df.drop('recorded_by', axis=1)"
      ],
      "execution_count": 326,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FOzWNiK4uNBw"
      },
      "source": [
        "# extraction_type, extraction_type_group and extraction_type_class appear to contain very similar\n",
        "# data. I'll drop the first two and keep the last one.\n",
        "\n",
        "train_df = train_df.drop(['extraction_type', 'extraction_type_group'], axis=1)"
      ],
      "execution_count": 327,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9JdgT13yuacT",
        "outputId": "7b3f3c11-0ea4-4753-d35a-2bb514679758"
      },
      "source": [
        "train_df.management.value_counts()"
      ],
      "execution_count": 328,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "vwc                 30371\n",
              "wug                  4874\n",
              "water board          2221\n",
              "wua                  1907\n",
              "private operator     1485\n",
              "parastatal           1347\n",
              "water authority       677\n",
              "other                 622\n",
              "company               512\n",
              "unknown               402\n",
              "other - school         70\n",
              "trust                  62\n",
              "Name: management, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 328
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4w9KqbW6uiQy"
      },
      "source": [
        "# This appears to be almost identical to 'scheme_management'. I'll drop it.\n",
        "\n",
        "train_df = train_df.drop('management', axis=1)"
      ],
      "execution_count": 329,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "l0WZWBcWu6Wo",
        "outputId": "338825b4-342a-4a97-97ea-f9efc5fc65aa"
      },
      "source": [
        "train_df.management_group.value_counts()"
      ],
      "execution_count": 331,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "user-group    39373\n",
              "commercial     2736\n",
              "parastatal     1347\n",
              "other           692\n",
              "unknown         402\n",
              "Name: management_group, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 331
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tFd3Yd7Pu-YX"
      },
      "source": [
        "# Appears to offer no new info and is likely to overlap with 'scheme_management'.\n",
        "\n",
        "train_df = train_df.drop('management_group', axis=1)"
      ],
      "execution_count": 332,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "y4rbJYF1vJa_",
        "outputId": "d34c906d-1e41-44ff-ff3e-0b10f74c7d95"
      },
      "source": [
        "train_df.payment.value_counts()"
      ],
      "execution_count": 333,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "never pay                19002\n",
              "pay per bucket            6808\n",
              "pay monthly               6182\n",
              "unknown                   6104\n",
              "pay when scheme fails     2941\n",
              "pay annually              2717\n",
              "other                      796\n",
              "Name: payment, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 333
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DmROJfOhvNdU",
        "outputId": "6af37526-9367-4c30-b23d-a21f9bb60058"
      },
      "source": [
        "train_df.payment_type.value_counts()"
      ],
      "execution_count": 334,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "never pay     19002\n",
              "per bucket     6808\n",
              "monthly        6182\n",
              "unknown        6104\n",
              "on failure     2941\n",
              "annually       2717\n",
              "other           796\n",
              "Name: payment_type, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 334
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kdOIwN-XvWUb"
      },
      "source": [
        "# Payment and payment_type contain identical data. Remove one and keep the other.\n",
        "\n",
        "train_df = train_df.drop('payment', axis=1)"
      ],
      "execution_count": 335,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pRhxuvHCvgew",
        "outputId": "ba82e9d0-3a9e-4a78-8356-526e4404f369"
      },
      "source": [
        "train_df.water_quality.value_counts()"
      ],
      "execution_count": 336,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "soft                  38093\n",
              "salty                  3668\n",
              "unknown                1397\n",
              "milky                   606\n",
              "coloured                364\n",
              "salty abandoned         264\n",
              "fluoride                147\n",
              "fluoride abandoned       11\n",
              "Name: water_quality, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 336
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yrXo74TLvppd",
        "outputId": "3b3c029f-accd-40f5-a545-5fbade280634"
      },
      "source": [
        "train_df.quality_group.value_counts()"
      ],
      "execution_count": 337,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "good        38093\n",
              "salty        3932\n",
              "unknown      1397\n",
              "milky         606\n",
              "colored       364\n",
              "fluoride      158\n",
              "Name: quality_group, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 337
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qAi8iTCovuEf"
      },
      "source": [
        "# Water_quality and quality_group contain identical data. Remove one and keep the other.\n",
        "\n",
        "train_df = train_df.drop('quality_group', 1)"
      ],
      "execution_count": 338,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2owySXAGv0QR",
        "outputId": "2af5a737-1d0a-4ec7-979e-afd581a7bfd9"
      },
      "source": [
        "train_df.quantity.value_counts()"
      ],
      "execution_count": 339,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "enough          24821\n",
              "insufficient    11367\n",
              "dry              4725\n",
              "seasonal         3049\n",
              "unknown           588\n",
              "Name: quantity, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 339
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZNesaWR5v5lx",
        "outputId": "0c6d87b4-96b0-414c-fa12-56c2b36a34c7"
      },
      "source": [
        "train_df.quantity_group.value_counts()"
      ],
      "execution_count": 340,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "enough          24821\n",
              "insufficient    11367\n",
              "dry              4725\n",
              "seasonal         3049\n",
              "unknown           588\n",
              "Name: quantity_group, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 340
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "69F8C5Duv9U_"
      },
      "source": [
        "# Quantity and quantity_group contain identical data. Remove one and keep the other.\n",
        "\n",
        "train_df = train_df.drop('quantity_group', 1)"
      ],
      "execution_count": 341,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "35bNXYKUwDJt",
        "outputId": "486f445f-b38a-4be7-df16-e886aa476fb0"
      },
      "source": [
        "train_df.source.value_counts()"
      ],
      "execution_count": 342,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "spring                  12889\n",
              "shallow well            12618\n",
              "machine dbh              8282\n",
              "river                    7135\n",
              "rainwater harvesting     1739\n",
              "hand dtw                  652\n",
              "lake                      544\n",
              "dam                       489\n",
              "other                     156\n",
              "unknown                    46\n",
              "Name: source, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 342
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RAIJGHcGwKQS",
        "outputId": "9acbfa2e-c4dd-4c7c-e084-67b2ec7a9354"
      },
      "source": [
        "train_df.source_class.value_counts()"
      ],
      "execution_count": 343,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "groundwater    34441\n",
              "surface         9907\n",
              "unknown          202\n",
              "Name: source_class, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 343
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YM0aInYfwVrJ",
        "outputId": "8a4da58f-b7b5-480c-98ef-2311dc4a05b2"
      },
      "source": [
        "train_df.source_type.value_counts()"
      ],
      "execution_count": 344,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "spring                  12889\n",
              "shallow well            12618\n",
              "borehole                 8934\n",
              "river/lake               7679\n",
              "rainwater harvesting     1739\n",
              "dam                       489\n",
              "other                     202\n",
              "Name: source_type, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 344
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fNcYryEJwgFE"
      },
      "source": [
        "# Source and source_type contain very similar information. Remove one and keep the other.\n",
        "\n",
        "train_df = train_df.drop('source', axis=1)"
      ],
      "execution_count": 345,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R-O7I7qEwu4K"
      },
      "source": [
        "# gps_height, longitude, latitude, region_code and district_code are all geographic info which\n",
        "# is unlikely to add any predictive power to the model given that there are other variables\n",
        "# containing geographic data. We expect id to not contain any useful information so that gets\n",
        "# dropped too.\n",
        "\n",
        "train_df = train_df.drop(['gps_height', 'longitude', 'latitude', 'region_code', 'district_code', 'id'], axis=1)"
      ],
      "execution_count": 350,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E3K_Vjay1I2x"
      },
      "source": [
        "train_df = train_df.drop('num_private',axis=1)"
      ],
      "execution_count": 372,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ny0J3uhzxR0C",
        "outputId": "62824fa2-f457-46db-8103-8b58a9eec957"
      },
      "source": [
        "pd.set_option('display.max_rows', None)\n",
        "print(str_cols.apply(lambda x: len(x.unique())))\n",
        "pd.set_option('display.max_rows', 15)"
      ],
      "execution_count": 351,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "date_recorded              346\n",
            "funder                       6\n",
            "installer                    6\n",
            "wpt_name                 28991\n",
            "basin                        9\n",
            "region                      21\n",
            "lga                        125\n",
            "ward                      2080\n",
            "public_meeting               3\n",
            "recorded_by                  1\n",
            "scheme_management            6\n",
            "permit                       3\n",
            "extraction_type             18\n",
            "extraction_type_group       13\n",
            "extraction_type_class        7\n",
            "management                  12\n",
            "management_group             5\n",
            "payment                      7\n",
            "payment_type                 7\n",
            "water_quality                8\n",
            "quality_group                6\n",
            "quantity                     5\n",
            "quantity_group               5\n",
            "source                      10\n",
            "source_type                  7\n",
            "source_class                 3\n",
            "waterpoint_type              7\n",
            "waterpoint_type_group        6\n",
            "status_group                 3\n",
            "dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vofx-Mh_xW0-",
        "outputId": "e281d85e-4823-4dcd-c672-32f71dd3137c"
      },
      "source": [
        "train_df.construction_year.value_counts()"
      ],
      "execution_count": 352,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1950    15641\n",
              "2010     1991\n",
              "2008     1935\n",
              "2009     1928\n",
              "2000     1575\n",
              "        ...  \n",
              "1964       31\n",
              "1962       28\n",
              "1961       17\n",
              "1966       14\n",
              "1965       11\n",
              "Name: construction_year, Length: 55, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 352
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iVAGJ-P5xvw2"
      },
      "source": [
        "# Turn construction_year into a categorical column containing the following values: '60s', '70s',\n",
        "# '80s', '90s, '00s', '10s', 'unknown'.\n",
        "\n",
        "def construction_wrangler(row):\n",
        "  if row['construction_year'] >= 1960 and row['construction_year'] < 1970:\n",
        "    return '60s'\n",
        "  elif row['construction_year'] >= 1970 and row['construction_year'] < 1980:\n",
        "    return '70s'\n",
        "  elif row['construction_year'] >= 1980 and row['construction_year'] < 1990:\n",
        "    return '80s'\n",
        "  elif row['construction_year'] >= 1990 and row['construction_year'] < 2000:\n",
        "    return '90s'\n",
        "  elif row['construction_year'] >= 2000 and row['construction_year'] < 2010:\n",
        "    return '00s'\n",
        "  elif row['construction_year'] >= 2010:\n",
        "    return '10s'\n",
        "  else:\n",
        "    return 'unknown'\n",
        "    \n",
        "train_df['construction_year'] = train_df.apply(lambda row: construction_wrangler(row), axis=1)"
      ],
      "execution_count": 353,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "FcwZL2iCyGce",
        "outputId": "dc125e86-bdeb-42ba-eb9f-aacf186be46f"
      },
      "source": [
        "sns.histplot(train_df.population, bins = 40)\n",
        "plt.show()"
      ],
      "execution_count": 356,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rMfhlbfeyMGO",
        "outputId": "465a39f1-3a68-43d5-c0ee-eb7cc7934f77"
      },
      "source": [
        "train_df.info()"
      ],
      "execution_count": 357,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 44550 entries, 0 to 44549\n",
            "Data columns (total 21 columns):\n",
            " #   Column                 Non-Null Count  Dtype  \n",
            "---  ------                 --------------  -----  \n",
            " 0   amount_tsh             44550 non-null  float64\n",
            " 1   days_since_recorded    44550 non-null  int64  \n",
            " 2   funder                 44550 non-null  object \n",
            " 3   installer              44550 non-null  object \n",
            " 4   num_private            44550 non-null  int64  \n",
            " 5   basin                  44550 non-null  object \n",
            " 6   population             44550 non-null  int64  \n",
            " 7   public_meeting         44550 non-null  object \n",
            " 8   scheme_management      44550 non-null  object \n",
            " 9   permit                 44550 non-null  object \n",
            " 10  construction_year      44550 non-null  object \n",
            " 11  extraction_type_class  44550 non-null  object \n",
            " 12  payment_type           44550 non-null  object \n",
            " 13  water_quality          44550 non-null  object \n",
            " 14  quantity               44550 non-null  object \n",
            " 15  source_type            44550 non-null  object \n",
            " 16  source_class           44550 non-null  object \n",
            " 17  waterpoint_type        44550 non-null  object \n",
            " 18  waterpoint_type_group  44550 non-null  object \n",
            " 19  status_group           44550 non-null  object \n",
            " 20  status_group_vals      44550 non-null  int64  \n",
            "dtypes: float64(1), int64(4), object(16)\n",
            "memory usage: 7.1+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 280
        },
        "id": "O4728LNiyeK-",
        "outputId": "361e3a3b-839d-477a-9593-a2658236baa5"
      },
      "source": [
        "# Most wells have a few hundred people living around them. There are some wells\n",
        "# serving huge populations. This may skew the data.\n",
        "sns.histplot(train_df.amount_tsh, bins = 40)\n",
        "plt.show()"
      ],
      "execution_count": 360,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PIWWL0E-yzfX",
        "outputId": "2797a005-88e4-42cc-c0bb-96abf0029f61"
      },
      "source": [
        "# This plot measures the amount of water available at the pump. It looks a lot like the \n",
        "# population graph which makes sense. \n",
        "\n",
        "train_df.population.describe()"
      ],
      "execution_count": 361,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count    44550.000000\n",
              "mean       179.112750\n",
              "std        478.995645\n",
              "min          0.000000\n",
              "25%          0.000000\n",
              "50%         25.000000\n",
              "75%        210.000000\n",
              "max      30500.000000\n",
              "Name: population, dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 361
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CksUzZkHy53d",
        "outputId": "1a9fba1f-36b1-4a2e-f1b0-a308a21530e1"
      },
      "source": [
        "train_df.amount_tsh.describe()"
      ],
      "execution_count": 362,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count     44550.000000\n",
              "mean        307.623224\n",
              "std        2402.742799\n",
              "min           0.000000\n",
              "25%           0.000000\n",
              "50%           0.000000\n",
              "75%          20.000000\n",
              "max      170000.000000\n",
              "Name: amount_tsh, dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 362
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YzpYJDrly_Sn"
      },
      "source": [
        "# There appears to be enough variation between the two to warrant keeping them in the model.\n",
        "# Let's save the dataframe to a new csv file. We'll start creating models in the next notebooks.\n",
        "train_df = train_df.drop('status_group_vals', 1)\n",
        "train_df.to_csv('pump_train_for_models.csv', index=False)"
      ],
      "execution_count": 363,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c1X6q2vXzRN1"
      },
      "source": [
        "# We'll also need to perform the same modifications to the test set.\n",
        "\n",
        "test_df = test_df.drop(['gps_height', 'longitude', 'latitude', 'region_code', 'district_code','num_private', 'id', 'payment', 'management_group', 'management', 'extraction_type', 'extraction_type_group', 'recorded_by','region', 'lga','ward', 'wpt_name', 'scheme_name', 'subvillage', 'quantity_group','quality_group', 'source'], axis=1)"
      ],
      "execution_count": 393,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UKyhtSeyzsTN",
        "outputId": "aabecd54-3005-417a-a7cb-e7e05f3dd354"
      },
      "source": [
        "test_df.date_recorded = pd.datetime(2013, 12, 3) - pd.to_datetime(test_df.date_recorded)\n",
        "test_df.columns = ['days_since_recorded' if x=='date_recorded' else x for x in test_df.columns]\n",
        "test_df.days_since_recorded = test_df.days_since_recorded.astype('timedelta64[D]').astype(int)"
      ],
      "execution_count": 394,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:1: FutureWarning: The pandas.datetime class is deprecated and will be removed from pandas in a future version. Import from datetime module instead.\n",
            "  \"\"\"Entry point for launching an IPython kernel.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZfokWoCVz61j"
      },
      "source": [
        "test_df.permit = test_df.permit.fillna('Unknown')\n",
        "test_df.public_meeting = test_df.public_meeting.fillna('Unknown')"
      ],
      "execution_count": 395,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tRamguz30GBN"
      },
      "source": [
        "test_df['scheme_management'] = test_df.apply(lambda row: scheme_wrangler(row), axis=1)\n",
        "test_df['construction_year'] = test_df.apply(lambda row: construction_wrangler(row), axis=1)\n",
        "test_df['installer'] = test_df.apply(lambda row: installer_wrangler(row), axis=1)\n",
        "test_df['funder'] = test_df.apply(lambda row: funder_wrangler(row), axis=1)"
      ],
      "execution_count": 396,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uW6ydstf0O3W",
        "outputId": "db9720d7-525c-4a10-8318-89bd7c5fd115"
      },
      "source": [
        "train_df.shape"
      ],
      "execution_count": 397,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(44550, 94)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 397
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 635
        },
        "id": "6HnOrfOO0cO7",
        "outputId": "05201de3-d353-4e87-9fb3-58a54a9a1640"
      },
      "source": [
        "pd.set_option('display.max_rows', 30)\n",
        "display(train_df)\n",
        "pd.set_option('display.max_rows', 15)"
      ],
      "execution_count": 371,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>days_since_recorded</th>\n",
              "      <th>funder</th>\n",
              "      <th>installer</th>\n",
              "      <th>num_private</th>\n",
              "      <th>basin</th>\n",
              "      <th>population</th>\n",
              "      <th>public_meeting</th>\n",
              "      <th>scheme_management</th>\n",
              "      <th>permit</th>\n",
              "      <th>construction_year</th>\n",
              "      <th>extraction_type_class</th>\n",
              "      <th>payment_type</th>\n",
              "      <th>water_quality</th>\n",
              "      <th>quantity</th>\n",
              "      <th>source_type</th>\n",
              "      <th>source_class</th>\n",
              "      <th>waterpoint_type</th>\n",
              "      <th>waterpoint_type_group</th>\n",
              "      <th>status_group</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>7.0</td>\n",
              "      <td>258</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>90</td>\n",
              "      <td>True</td>\n",
              "      <td>wtr_brd</td>\n",
              "      <td>True</td>\n",
              "      <td>00s</td>\n",
              "      <td>gravity</td>\n",
              "      <td>per bucket</td>\n",
              "      <td>soft</td>\n",
              "      <td>enough</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.0</td>\n",
              "      <td>1016</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Nyasa</td>\n",
              "      <td>70</td>\n",
              "      <td>False</td>\n",
              "      <td>other</td>\n",
              "      <td>False</td>\n",
              "      <td>80s</td>\n",
              "      <td>gravity</td>\n",
              "      <td>never pay</td>\n",
              "      <td>soft</td>\n",
              "      <td>enough</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.0</td>\n",
              "      <td>333</td>\n",
              "      <td>gov</td>\n",
              "      <td>dwe</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>150</td>\n",
              "      <td>True</td>\n",
              "      <td>vwc</td>\n",
              "      <td>True</td>\n",
              "      <td>00s</td>\n",
              "      <td>gravity</td>\n",
              "      <td>never pay</td>\n",
              "      <td>soft</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>non functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.0</td>\n",
              "      <td>975</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Rufiji</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>vwc</td>\n",
              "      <td>True</td>\n",
              "      <td>unknown</td>\n",
              "      <td>gravity</td>\n",
              "      <td>annually</td>\n",
              "      <td>soft</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>river/lake</td>\n",
              "      <td>surface</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>300.0</td>\n",
              "      <td>270</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>34</td>\n",
              "      <td>True</td>\n",
              "      <td>vwc</td>\n",
              "      <td>False</td>\n",
              "      <td>00s</td>\n",
              "      <td>gravity</td>\n",
              "      <td>monthly</td>\n",
              "      <td>soft</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44545</th>\n",
              "      <td>0.0</td>\n",
              "      <td>314</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Tanganyika</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>other</td>\n",
              "      <td>True</td>\n",
              "      <td>unknown</td>\n",
              "      <td>handpump</td>\n",
              "      <td>unknown</td>\n",
              "      <td>milky</td>\n",
              "      <td>enough</td>\n",
              "      <td>shallow well</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44546</th>\n",
              "      <td>0.0</td>\n",
              "      <td>862</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Victoria</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>wua</td>\n",
              "      <td>True</td>\n",
              "      <td>unknown</td>\n",
              "      <td>handpump</td>\n",
              "      <td>monthly</td>\n",
              "      <td>soft</td>\n",
              "      <td>enough</td>\n",
              "      <td>borehole</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>non functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44547</th>\n",
              "      <td>0.0</td>\n",
              "      <td>283</td>\n",
              "      <td>gov</td>\n",
              "      <td>gov</td>\n",
              "      <td>0</td>\n",
              "      <td>Ruvuma / Southern Coast</td>\n",
              "      <td>60</td>\n",
              "      <td>Unknown</td>\n",
              "      <td>vwc</td>\n",
              "      <td>Unknown</td>\n",
              "      <td>70s</td>\n",
              "      <td>gravity</td>\n",
              "      <td>unknown</td>\n",
              "      <td>soft</td>\n",
              "      <td>seasonal</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44548</th>\n",
              "      <td>200.0</td>\n",
              "      <td>252</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Pangani</td>\n",
              "      <td>80</td>\n",
              "      <td>True</td>\n",
              "      <td>wtr_auth</td>\n",
              "      <td>True</td>\n",
              "      <td>10s</td>\n",
              "      <td>gravity</td>\n",
              "      <td>monthly</td>\n",
              "      <td>soft</td>\n",
              "      <td>insufficient</td>\n",
              "      <td>spring</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>communal standpipe</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44549</th>\n",
              "      <td>0.0</td>\n",
              "      <td>311</td>\n",
              "      <td>other</td>\n",
              "      <td>other</td>\n",
              "      <td>0</td>\n",
              "      <td>Lake Tanganyika</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>vwc</td>\n",
              "      <td>False</td>\n",
              "      <td>unknown</td>\n",
              "      <td>handpump</td>\n",
              "      <td>on failure</td>\n",
              "      <td>soft</td>\n",
              "      <td>enough</td>\n",
              "      <td>shallow well</td>\n",
              "      <td>groundwater</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>hand pump</td>\n",
              "      <td>functional</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>44550 rows × 20 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       amount_tsh  days_since_recorded  ... waterpoint_type_group    status_group\n",
              "0             7.0                  258  ...    communal standpipe      functional\n",
              "1             0.0                 1016  ...    communal standpipe      functional\n",
              "2             0.0                  333  ...    communal standpipe  non functional\n",
              "3             0.0                  975  ...    communal standpipe      functional\n",
              "4           300.0                  270  ...    communal standpipe      functional\n",
              "...           ...                  ...  ...                   ...             ...\n",
              "44545         0.0                  314  ...             hand pump      functional\n",
              "44546         0.0                  862  ...             hand pump  non functional\n",
              "44547         0.0                  283  ...    communal standpipe      functional\n",
              "44548       200.0                  252  ...    communal standpipe      functional\n",
              "44549         0.0                  311  ...             hand pump      functional\n",
              "\n",
              "[44550 rows x 20 columns]"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2gnNtNBG0TQu",
        "outputId": "6a74aea9-f265-472b-876f-a342d6c89ded"
      },
      "source": [
        "test_df.shape"
      ],
      "execution_count": 369,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(14850, 19)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 369
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T6s_Kb6N0XU4"
      },
      "source": [
        "# The train and test sets match up. We can save the test set now.\n",
        "\n",
        "test_df.to_csv('pump_test_for_models.csv', index=False)"
      ],
      "execution_count": 373,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "seT5L7U4S_IU"
      },
      "source": [
        "# ML Models & Prediction"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pp69nQPZIzbc"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.svm import SVC\n",
        "from sklearn.ensemble import BaggingClassifier\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.preprocessing import StandardScaler"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dtpgd3X1JEkW",
        "outputId": "20ad2a6c-c627-497e-b957-0503f02d2107"
      },
      "source": [
        "%rm -rf ML-Assignment-data\n",
        "!git clone https://github.com/py-here/ML-Assignment-data"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'ML-Assignment-data'...\n",
            "remote: Enumerating objects: 13, done.\u001b[K\n",
            "remote: Counting objects: 100% (13/13), done.\u001b[K\n",
            "remote: Compressing objects: 100% (11/11), done.\u001b[K\n",
            "remote: Total 13 (delta 3), reused 0 (delta 0), pack-reused 0\u001b[K\n",
            "Unpacking objects: 100% (13/13), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yDAmCbiJK0im"
      },
      "source": [
        "train_df = pd.read_csv(\"/content/ML-Assignment-data/pump_train_for_models.csv\")\n",
        "test_df = pd.read_csv(\"/content/ML-Assignment-data/pump_test_for_models.csv\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Il2mUJMl1oR6",
        "outputId": "43af64de-47ca-4da7-ddc1-e22dd07db133"
      },
      "source": [
        "train_df.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 44550 entries, 0 to 44549\n",
            "Data columns (total 20 columns):\n",
            " #   Column                 Non-Null Count  Dtype  \n",
            "---  ------                 --------------  -----  \n",
            " 0   amount_tsh             44550 non-null  float64\n",
            " 1   days_since_recorded    44550 non-null  int64  \n",
            " 2   funder                 44550 non-null  object \n",
            " 3   installer              44550 non-null  object \n",
            " 4   num_private            44550 non-null  int64  \n",
            " 5   basin                  44550 non-null  object \n",
            " 6   population             44550 non-null  int64  \n",
            " 7   public_meeting         44550 non-null  object \n",
            " 8   scheme_management      44550 non-null  object \n",
            " 9   permit                 44550 non-null  object \n",
            " 10  construction_year      44550 non-null  object \n",
            " 11  extraction_type_class  44550 non-null  object \n",
            " 12  payment_type           44550 non-null  object \n",
            " 13  water_quality          44550 non-null  object \n",
            " 14  quantity               44550 non-null  object \n",
            " 15  source_type            44550 non-null  object \n",
            " 16  source_class           44550 non-null  object \n",
            " 17  waterpoint_type        44550 non-null  object \n",
            " 18  waterpoint_type_group  44550 non-null  object \n",
            " 19  status_group           44550 non-null  object \n",
            "dtypes: float64(1), int64(3), object(16)\n",
            "memory usage: 6.8+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wt5ShdXc3Fvm"
      },
      "source": [
        "# Get dummy columns for the categorical columns and shuffle the data.\n",
        "\n",
        "dummy_cols = ['funder', 'installer', 'basin', 'public_meeting', 'scheme_management', 'permit',\n",
        "              'construction_year', 'extraction_type_class', 'payment_type', 'water_quality',\n",
        "              'quantity', 'source_type', 'source_class', 'waterpoint_type',\n",
        "              'waterpoint_type_group']\n",
        "\n",
        "train_df = pd.get_dummies(train_df, columns = dummy_cols)\n",
        "\n",
        "train_df = train_df.sample(frac=1).reset_index(drop=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 326
        },
        "id": "N1QnJVrA3ZV5",
        "outputId": "273882aa-84bb-40b5-b1f0-455b0ddbc449"
      },
      "source": [
        "print(train_df.shape)\n",
        "train_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(44550, 95)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>days_since_recorded</th>\n",
              "      <th>num_private</th>\n",
              "      <th>population</th>\n",
              "      <th>status_group</th>\n",
              "      <th>funder_danida</th>\n",
              "      <th>funder_gov</th>\n",
              "      <th>funder_hesawa</th>\n",
              "      <th>funder_other</th>\n",
              "      <th>funder_rwssp</th>\n",
              "      <th>funder_world_bank</th>\n",
              "      <th>installer_commu</th>\n",
              "      <th>installer_danida</th>\n",
              "      <th>installer_dwe</th>\n",
              "      <th>installer_gov</th>\n",
              "      <th>installer_other</th>\n",
              "      <th>installer_rwe</th>\n",
              "      <th>basin_Internal</th>\n",
              "      <th>basin_Lake Nyasa</th>\n",
              "      <th>basin_Lake Rukwa</th>\n",
              "      <th>basin_Lake Tanganyika</th>\n",
              "      <th>basin_Lake Victoria</th>\n",
              "      <th>basin_Pangani</th>\n",
              "      <th>basin_Rufiji</th>\n",
              "      <th>basin_Ruvuma / Southern Coast</th>\n",
              "      <th>basin_Wami / Ruvu</th>\n",
              "      <th>public_meeting_False</th>\n",
              "      <th>public_meeting_True</th>\n",
              "      <th>public_meeting_Unknown</th>\n",
              "      <th>scheme_management_other</th>\n",
              "      <th>scheme_management_vwc</th>\n",
              "      <th>scheme_management_wtr_auth</th>\n",
              "      <th>scheme_management_wtr_brd</th>\n",
              "      <th>scheme_management_wua</th>\n",
              "      <th>scheme_management_wug</th>\n",
              "      <th>permit_False</th>\n",
              "      <th>permit_True</th>\n",
              "      <th>permit_Unknown</th>\n",
              "      <th>construction_year_00s</th>\n",
              "      <th>construction_year_10s</th>\n",
              "      <th>...</th>\n",
              "      <th>payment_type_on failure</th>\n",
              "      <th>payment_type_other</th>\n",
              "      <th>payment_type_per bucket</th>\n",
              "      <th>payment_type_unknown</th>\n",
              "      <th>water_quality_coloured</th>\n",
              "      <th>water_quality_fluoride</th>\n",
              "      <th>water_quality_fluoride abandoned</th>\n",
              "      <th>water_quality_milky</th>\n",
              "      <th>water_quality_salty</th>\n",
              "      <th>water_quality_salty abandoned</th>\n",
              "      <th>water_quality_soft</th>\n",
              "      <th>water_quality_unknown</th>\n",
              "      <th>quantity_dry</th>\n",
              "      <th>quantity_enough</th>\n",
              "      <th>quantity_insufficient</th>\n",
              "      <th>quantity_seasonal</th>\n",
              "      <th>quantity_unknown</th>\n",
              "      <th>source_type_borehole</th>\n",
              "      <th>source_type_dam</th>\n",
              "      <th>source_type_other</th>\n",
              "      <th>source_type_rainwater harvesting</th>\n",
              "      <th>source_type_river/lake</th>\n",
              "      <th>source_type_shallow well</th>\n",
              "      <th>source_type_spring</th>\n",
              "      <th>source_class_groundwater</th>\n",
              "      <th>source_class_surface</th>\n",
              "      <th>source_class_unknown</th>\n",
              "      <th>waterpoint_type_cattle trough</th>\n",
              "      <th>waterpoint_type_communal standpipe</th>\n",
              "      <th>waterpoint_type_communal standpipe multiple</th>\n",
              "      <th>waterpoint_type_dam</th>\n",
              "      <th>waterpoint_type_hand pump</th>\n",
              "      <th>waterpoint_type_improved spring</th>\n",
              "      <th>waterpoint_type_other</th>\n",
              "      <th>waterpoint_type_group_cattle trough</th>\n",
              "      <th>waterpoint_type_group_communal standpipe</th>\n",
              "      <th>waterpoint_type_group_dam</th>\n",
              "      <th>waterpoint_type_group_hand pump</th>\n",
              "      <th>waterpoint_type_group_improved spring</th>\n",
              "      <th>waterpoint_type_group_other</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.0</td>\n",
              "      <td>281</td>\n",
              "      <td>0</td>\n",
              "      <td>400</td>\n",
              "      <td>non functional</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.0</td>\n",
              "      <td>272</td>\n",
              "      <td>0</td>\n",
              "      <td>203</td>\n",
              "      <td>functional</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>300.0</td>\n",
              "      <td>268</td>\n",
              "      <td>0</td>\n",
              "      <td>380</td>\n",
              "      <td>non functional</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.0</td>\n",
              "      <td>961</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>functional needs repair</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>50.0</td>\n",
              "      <td>265</td>\n",
              "      <td>0</td>\n",
              "      <td>150</td>\n",
              "      <td>functional</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 95 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   amount_tsh  ...  waterpoint_type_group_other\n",
              "0         0.0  ...                            0\n",
              "1         0.0  ...                            0\n",
              "2       300.0  ...                            1\n",
              "3         0.0  ...                            0\n",
              "4        50.0  ...                            0\n",
              "\n",
              "[5 rows x 95 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z2U-4Gso3dv4"
      },
      "source": [
        "test_df = pd.get_dummies(test_df, columns = dummy_cols)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 309
        },
        "id": "6R_6aFuW3mOW",
        "outputId": "96fa3507-42e3-4105-b291-2aedf98093fe"
      },
      "source": [
        "print(test_df.shape)\n",
        "test_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(14850, 94)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>days_since_recorded</th>\n",
              "      <th>population</th>\n",
              "      <th>funder_danida</th>\n",
              "      <th>funder_gov</th>\n",
              "      <th>funder_hesawa</th>\n",
              "      <th>funder_other</th>\n",
              "      <th>funder_rwssp</th>\n",
              "      <th>funder_world_bank</th>\n",
              "      <th>installer_commu</th>\n",
              "      <th>installer_danida</th>\n",
              "      <th>installer_dwe</th>\n",
              "      <th>installer_gov</th>\n",
              "      <th>installer_other</th>\n",
              "      <th>installer_rwe</th>\n",
              "      <th>basin_Internal</th>\n",
              "      <th>basin_Lake Nyasa</th>\n",
              "      <th>basin_Lake Rukwa</th>\n",
              "      <th>basin_Lake Tanganyika</th>\n",
              "      <th>basin_Lake Victoria</th>\n",
              "      <th>basin_Pangani</th>\n",
              "      <th>basin_Rufiji</th>\n",
              "      <th>basin_Ruvuma / Southern Coast</th>\n",
              "      <th>basin_Wami / Ruvu</th>\n",
              "      <th>public_meeting_False</th>\n",
              "      <th>public_meeting_True</th>\n",
              "      <th>public_meeting_Unknown</th>\n",
              "      <th>scheme_management_other</th>\n",
              "      <th>scheme_management_vwc</th>\n",
              "      <th>scheme_management_wtr_auth</th>\n",
              "      <th>scheme_management_wtr_brd</th>\n",
              "      <th>scheme_management_wua</th>\n",
              "      <th>scheme_management_wug</th>\n",
              "      <th>permit_False</th>\n",
              "      <th>permit_True</th>\n",
              "      <th>permit_Unknown</th>\n",
              "      <th>construction_year_00s</th>\n",
              "      <th>construction_year_10s</th>\n",
              "      <th>construction_year_60s</th>\n",
              "      <th>...</th>\n",
              "      <th>payment_type_on failure</th>\n",
              "      <th>payment_type_other</th>\n",
              "      <th>payment_type_per bucket</th>\n",
              "      <th>payment_type_unknown</th>\n",
              "      <th>water_quality_coloured</th>\n",
              "      <th>water_quality_fluoride</th>\n",
              "      <th>water_quality_fluoride abandoned</th>\n",
              "      <th>water_quality_milky</th>\n",
              "      <th>water_quality_salty</th>\n",
              "      <th>water_quality_salty abandoned</th>\n",
              "      <th>water_quality_soft</th>\n",
              "      <th>water_quality_unknown</th>\n",
              "      <th>quantity_dry</th>\n",
              "      <th>quantity_enough</th>\n",
              "      <th>quantity_insufficient</th>\n",
              "      <th>quantity_seasonal</th>\n",
              "      <th>quantity_unknown</th>\n",
              "      <th>source_type_borehole</th>\n",
              "      <th>source_type_dam</th>\n",
              "      <th>source_type_other</th>\n",
              "      <th>source_type_rainwater harvesting</th>\n",
              "      <th>source_type_river/lake</th>\n",
              "      <th>source_type_shallow well</th>\n",
              "      <th>source_type_spring</th>\n",
              "      <th>source_class_groundwater</th>\n",
              "      <th>source_class_surface</th>\n",
              "      <th>source_class_unknown</th>\n",
              "      <th>waterpoint_type_cattle trough</th>\n",
              "      <th>waterpoint_type_communal standpipe</th>\n",
              "      <th>waterpoint_type_communal standpipe multiple</th>\n",
              "      <th>waterpoint_type_dam</th>\n",
              "      <th>waterpoint_type_hand pump</th>\n",
              "      <th>waterpoint_type_improved spring</th>\n",
              "      <th>waterpoint_type_other</th>\n",
              "      <th>waterpoint_type_group_cattle trough</th>\n",
              "      <th>waterpoint_type_group_communal standpipe</th>\n",
              "      <th>waterpoint_type_group_dam</th>\n",
              "      <th>waterpoint_type_group_hand pump</th>\n",
              "      <th>waterpoint_type_group_improved spring</th>\n",
              "      <th>waterpoint_type_group_other</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1012</td>\n",
              "      <td>769</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>961</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>982</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>0.0</td>\n",
              "      <td>873</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>0.0</td>\n",
              "      <td>856</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 94 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   Unnamed: 0  ...  waterpoint_type_group_other\n",
              "0           0  ...                            0\n",
              "1           1  ...                            0\n",
              "2           2  ...                            0\n",
              "3           3  ...                            0\n",
              "4           4  ...                            0\n",
              "\n",
              "[5 rows x 94 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 290
        },
        "id": "JC-0qH1E4VbX",
        "outputId": "749d3f1b-3412-4856-ac0c-80b06d92b128"
      },
      "source": [
        "test_df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>amount_tsh</th>\n",
              "      <th>days_since_recorded</th>\n",
              "      <th>population</th>\n",
              "      <th>funder_danida</th>\n",
              "      <th>funder_gov</th>\n",
              "      <th>funder_hesawa</th>\n",
              "      <th>funder_other</th>\n",
              "      <th>funder_rwssp</th>\n",
              "      <th>funder_world_bank</th>\n",
              "      <th>installer_commu</th>\n",
              "      <th>installer_danida</th>\n",
              "      <th>installer_dwe</th>\n",
              "      <th>installer_gov</th>\n",
              "      <th>installer_other</th>\n",
              "      <th>installer_rwe</th>\n",
              "      <th>basin_Internal</th>\n",
              "      <th>basin_Lake Nyasa</th>\n",
              "      <th>basin_Lake Rukwa</th>\n",
              "      <th>basin_Lake Tanganyika</th>\n",
              "      <th>basin_Lake Victoria</th>\n",
              "      <th>basin_Pangani</th>\n",
              "      <th>basin_Rufiji</th>\n",
              "      <th>basin_Ruvuma / Southern Coast</th>\n",
              "      <th>basin_Wami / Ruvu</th>\n",
              "      <th>public_meeting_False</th>\n",
              "      <th>public_meeting_True</th>\n",
              "      <th>public_meeting_Unknown</th>\n",
              "      <th>scheme_management_other</th>\n",
              "      <th>scheme_management_vwc</th>\n",
              "      <th>scheme_management_wtr_auth</th>\n",
              "      <th>scheme_management_wtr_brd</th>\n",
              "      <th>scheme_management_wua</th>\n",
              "      <th>scheme_management_wug</th>\n",
              "      <th>permit_False</th>\n",
              "      <th>permit_True</th>\n",
              "      <th>permit_Unknown</th>\n",
              "      <th>construction_year_00s</th>\n",
              "      <th>construction_year_10s</th>\n",
              "      <th>construction_year_60s</th>\n",
              "      <th>...</th>\n",
              "      <th>payment_type_on failure</th>\n",
              "      <th>payment_type_other</th>\n",
              "      <th>payment_type_per bucket</th>\n",
              "      <th>payment_type_unknown</th>\n",
              "      <th>water_quality_coloured</th>\n",
              "      <th>water_quality_fluoride</th>\n",
              "      <th>water_quality_fluoride abandoned</th>\n",
              "      <th>water_quality_milky</th>\n",
              "      <th>water_quality_salty</th>\n",
              "      <th>water_quality_salty abandoned</th>\n",
              "      <th>water_quality_soft</th>\n",
              "      <th>water_quality_unknown</th>\n",
              "      <th>quantity_dry</th>\n",
              "      <th>quantity_enough</th>\n",
              "      <th>quantity_insufficient</th>\n",
              "      <th>quantity_seasonal</th>\n",
              "      <th>quantity_unknown</th>\n",
              "      <th>source_type_borehole</th>\n",
              "      <th>source_type_dam</th>\n",
              "      <th>source_type_other</th>\n",
              "      <th>source_type_rainwater harvesting</th>\n",
              "      <th>source_type_river/lake</th>\n",
              "      <th>source_type_shallow well</th>\n",
              "      <th>source_type_spring</th>\n",
              "      <th>source_class_groundwater</th>\n",
              "      <th>source_class_surface</th>\n",
              "      <th>source_class_unknown</th>\n",
              "      <th>waterpoint_type_cattle trough</th>\n",
              "      <th>waterpoint_type_communal standpipe</th>\n",
              "      <th>waterpoint_type_communal standpipe multiple</th>\n",
              "      <th>waterpoint_type_dam</th>\n",
              "      <th>waterpoint_type_hand pump</th>\n",
              "      <th>waterpoint_type_improved spring</th>\n",
              "      <th>waterpoint_type_other</th>\n",
              "      <th>waterpoint_type_group_cattle trough</th>\n",
              "      <th>waterpoint_type_group_communal standpipe</th>\n",
              "      <th>waterpoint_type_group_dam</th>\n",
              "      <th>waterpoint_type_group_hand pump</th>\n",
              "      <th>waterpoint_type_group_improved spring</th>\n",
              "      <th>waterpoint_type_group_other</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1012</td>\n",
              "      <td>769</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>961</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>982</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>0.0</td>\n",
              "      <td>873</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>0.0</td>\n",
              "      <td>856</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 94 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   Unnamed: 0  ...  waterpoint_type_group_other\n",
              "0           0  ...                            0\n",
              "1           1  ...                            0\n",
              "2           2  ...                            0\n",
              "3           3  ...                            0\n",
              "4           4  ...                            0\n",
              "\n",
              "[5 rows x 94 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EgVXfMvg3t-U"
      },
      "source": [
        "# Let's split the train set into train and validation sets. Also remove the target.\n",
        "\n",
        "target = train_df.status_group\n",
        "features = train_df.drop('status_group', axis=1)\n",
        "\n",
        "X_train, X_val, y_train, y_val = train_test_split(features, target, train_size=0.8)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xHXPwB6TyPeD"
      },
      "source": [
        "#found from random forrest classifier\n",
        "major_features = ['amount_tsh','days_since_recorded','population','funder_gov',\n",
        "                  'funder_other','installer_dwe','installer_other','scheme_management_vwc',\n",
        "                  'permit_False','permit_True','construction_year_00s','construction_year_10s',\n",
        "                  'construction_year_70s','extraction_type_class_gravity','extraction_type_class_other',\n",
        "                  'payment_type_never pay','quantity_dry','quantity_enough','quantity_insufficient',\n",
        "                  'waterpoint_type_communal standpipe','waterpoint_type_communal standpipe multiple',\n",
        "                  'waterpoint_type_other','waterpoint_type_group_other']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0I1TftEsy768"
      },
      "source": [
        "new_X_train = X_train[major_features]\n",
        "new_X_val = X_val[major_features]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3PHYLzIJ-hN6"
      },
      "source": [
        "## Gradient Boosting"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PUM1bYh-vWea"
      },
      "source": [
        "def modelGB(X_train, X_val, y_train, y_val, test):\n",
        "  if __name__ == '__main__':\n",
        "   \n",
        "\n",
        "    param_grid = {'learning_rate': [0.075,0.07],'max_depth': [6,7],'min_samples_leaf': [7,8],\n",
        "                  'max_features': [1.0],'n_estimators':[100, 200]}  \n",
        "\n",
        "    estimator = GridSearchCV(estimator=GradientBoostingClassifier(),param_grid=param_grid,n_jobs=-1)\n",
        "\n",
        "    estimator.fit(X_train, y_train)\n",
        "\n",
        "    best_params = estimator.best_params_\n",
        "                              \n",
        "    validation_accuracy = estimator.score(X_val, y_val)\n",
        "    print('Validation accuracy: ', validation_accuracy)\n",
        "    print(best_params)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KsDLBeavwDo-",
        "outputId": "1c8c2201-d88d-4cd2-d10f-a90b5c37746e"
      },
      "source": [
        "#used reduced set of features as it was taking lot of time to compute\n",
        "modelGB(new_X_train, new_X_val, y_train, y_val, test_df)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Validation accuracy:  0.7734006734006734\n",
            "{'learning_rate': 0.075, 'max_depth': 7, 'max_features': 1.0, 'min_samples_leaf': 7, 'n_estimators': 200}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pRJHz1Nx7FnH"
      },
      "source": [
        "    Validation accuracy:  0.7734006734006734\n",
        "    {'learning_rate': 0.075, 'max_depth': 7, 'max_features': 1.0, 'min_samples_leaf': 7, 'n_estimators': 200}"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J1a9qmB9wFlE"
      },
      "source": [
        "def predictGB(features, target, test):\n",
        "  if __name__ == '__main__':\n",
        "\n",
        "    param_grid = {'learning_rate': [0.075],'max_depth': [7],'min_samples_leaf': [7],\n",
        "                  'max_features': [1.0],'n_estimators':[200]}  \n",
        "    estimator = GridSearchCV(estimator=GradientBoostingClassifier(),param_grid=param_grid)\n",
        "    estimator.fit(features, target)        \n",
        "\n",
        "    predictions = estimator.predict(test)\n",
        "    data = { 'status_group': predictions}\n",
        "\n",
        "    result = pd.DataFrame(data=data)\n",
        "\n",
        "    vals_to_replace = {2:'functional', 1:'functional needs repair', 0:'non functional'}\n",
        "    result.status_group = result.status_group.replace(vals_to_replace)        \n",
        "\n",
        "    result.to_csv('pump_predictions_GB.csv', index=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HkA-TZLzwZDq"
      },
      "source": [
        "# Run model for prediction\n",
        "\n",
        "predictGB(features, target, test_df)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9wLdDB90THwc"
      },
      "source": [
        "## Linear SVC"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kkbYzk9r5dk7"
      },
      "source": [
        "Both the train and test set are ready for modelling. Start off with a Linear Support Vector Classification.\n",
        "\n",
        "Ref Link: http://scikit-learn.org/stable/modules/generated/sklearn.svm.LinearSVC.html"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_jkmF3VbDYkg"
      },
      "source": [
        "def modelSVC(X_train, X_val, y_train, y_val, test):\n",
        "  if __name__ == '__main__':\n",
        "\n",
        "    pipe = Pipeline([('scl', StandardScaler()), ('clf', LinearSVC())])\n",
        "\n",
        "    param_grid = {'clf__C':[0.01, 0.1],'clf__class_weight':[None,'balanced']}\n",
        "\n",
        "    estimator = GridSearchCV(estimator=pipe,param_grid=param_grid,n_jobs=-1)\n",
        "\n",
        "    estimator.fit(X_train, y_train)\n",
        "\n",
        "    best_params = estimator.best_params_\n",
        "                              \n",
        "    validation_accuracy = estimator.score(X_val, y_val)\n",
        "    print('Validation accuracy: ', validation_accuracy)\n",
        "    print(best_params)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xd_qSTAlDYkh",
        "outputId": "4d1b2519-c083-470c-a75d-2b47846e1827"
      },
      "source": [
        "modelSVC(X_train, X_val, y_train, y_val, test_df)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Validation accuracy:  0.7306397306397306\n",
            "{'clf__C': 0.01, 'clf__class_weight': None}\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xlEDuZ7yDYki"
      },
      "source": [
        "    Validation accuracy:  0.7306397306397306\n",
        "    {'clf__C': 0.01, 'clf__class_weight': None}"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K86PQEABDYki"
      },
      "source": [
        "def predictSVC(features, target, test):\n",
        "  if __name__ == '__main__':\n",
        "\n",
        "    pipe = Pipeline([('scl', StandardScaler()), ('clf', LinearSVC())])\n",
        "\n",
        "    param_grid = {'clf__C':[0.01],'clf__class_weight':[None]}\n",
        "\n",
        "    estimator = GridSearchCV(estimator=pipe,param_grid=param_grid)\n",
        "\n",
        "    estimator.fit(features, target)        \n",
        "\n",
        "    predictions = estimator.predict(test)\n",
        "    data = { 'status_group': predictions}\n",
        "\n",
        "    result = pd.DataFrame(data=data)\n",
        "\n",
        "    vals_to_replace = {2:'functional', 1:'functional needs repair', 0:'non functional'}\n",
        "    result.status_group = result.status_group.replace(vals_to_replace)        \n",
        "\n",
        "    result.to_csv('pump_predictions.csv', index=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2IcqaegADYki",
        "outputId": "e164de59-5bcb-40ec-f242-c6c9524b3f3b"
      },
      "source": [
        "# Run model for prediction\n",
        "\n",
        "predictSVC(features, target, test_df)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/svm/_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LSdne2tiTLtG"
      },
      "source": [
        "## Bagging"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "diVAO86CNKTV"
      },
      "source": [
        "def modelBagging(X_train, X_val, y_train, y_val, test):\n",
        "  if __name__ == '__main__':\n",
        "\n",
        "    param_grid = {'n_estimators':[50,100,150],'max_samples':[1,5,10]}\n",
        "\n",
        "    estimator = GridSearchCV(estimator=BaggingClassifier(),param_grid=param_grid)\n",
        "\n",
        "    estimator.fit(X_train, y_train)\n",
        "\n",
        "    best_params = estimator.best_params_\n",
        "                              \n",
        "    validation_accuracy = estimator.score(X_val, y_val)\n",
        "    print('Validation accuracy: ', validation_accuracy)\n",
        "    print(best_params)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ftVhl-3iNsda",
        "outputId": "dac615b9-05a6-4b20-9e59-53fe47e2e81a"
      },
      "source": [
        "modelBagging(X_train, X_val, y_train, y_val,test_df)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Validation accuracy:  0.6668911335578003\n",
            "{'max_samples': 10, 'n_estimators': 100}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lpbTJnUkPkZ6"
      },
      "source": [
        "    Validation accuracy:  0.6515151515151515\n",
        "    {'max_samples': 10, 'n_estimators': 150}"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kr0-nHHJANBe"
      },
      "source": [
        "def predictBagging(features, target, test):\n",
        "  if __name__ == '__main__':\n",
        "\n",
        "    param_grid = {'n_estimators': [150],'max_samples': [10]}\n",
        "\n",
        "    estimator = GridSearchCV(estimator=BaggingClassifier(),param_grid=param_grid,n_jobs=-1)\n",
        "    estimator.fit(features, target)        \n",
        "\n",
        "    predictions = estimator.predict(test)\n",
        "    data = { 'status_group': predictions}\n",
        "\n",
        "    result = pd.DataFrame(data=data)\n",
        "\n",
        "    vals_to_replace = {2:'functional', 1:'functional needs repair', 0:'non functional'}\n",
        "    result.status_group = result.status_group.replace(vals_to_replace)        \n",
        "\n",
        "    result.to_csv('pump_predictions.csv', index=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "edyAvK1rBhKn"
      },
      "source": [
        "# Run model for prediction\n",
        "\n",
        "predictBagging(features, target, test_df)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VQku5nzVTPKG"
      },
      "source": [
        "## Decision Tree"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZDOL-K2_Q-ch"
      },
      "source": [
        "def modelDT(X_train, X_val, y_train, y_val, test):\n",
        "  if __name__ == '__main__':\n",
        "  \n",
        "    param_grid = {'max_depth':[1,5,10,20,30]}                      \n",
        "\n",
        "    estimator = GridSearchCV(estimator=DecisionTreeClassifier(),param_grid=param_grid, n_jobs=-1)\n",
        "\n",
        "    estimator.fit(X_train, y_train)\n",
        "\n",
        "    best_params = estimator.best_params_\n",
        "\n",
        "    print (best_params)\n",
        "                              \n",
        "    validation_accuracy = estimator.score(X_val, y_val)\n",
        "    print('Validation accuracy: ', validation_accuracy)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HmSKH28JRw08",
        "outputId": "10d9f8fb-5423-43bd-8815-818a71f7d363"
      },
      "source": [
        "modelDT(X_train, X_val, y_train, y_val, test_df)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{'max_depth': 20}\n",
            "Validation accuracy:  0.7594837261503928\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zhGVWu5PR1UV"
      },
      "source": [
        "def predictDT(features, target, test):\n",
        "  if __name__ == '__main__':\n",
        "\n",
        "    best_params = {'max_depth':[20]}                      \n",
        "\n",
        "    estimator = GridSearchCV(estimator=DecisionTreeClassifier(),param_grid=best_params,n_jobs=-1)\n",
        "    estimator.fit(features, target)     \n",
        "\n",
        "    predictions = estimator.predict(test)\n",
        "\n",
        "    data = {'status_group': predictions}\n",
        "    result = pd.DataFrame(data=data)\n",
        "    vals_to_replace = {2:'functional', 1:'functional needs repair',\n",
        "                      0:'non functional'}\n",
        "    result.status_group = result.status_group.replace(vals_to_replace)        \n",
        "    result.to_csv('pump_predictions.csv', index=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VzUc9ce8ZvWT"
      },
      "source": [
        "# Run model for prediction\n",
        "\n",
        "predictDT(features, target, test_df)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sDYrg3Xhov6C"
      },
      "source": [
        "# Feature Selection"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FHArLntxuNau"
      },
      "source": [
        "Here we are using Random Forest Classifier to identify relative feature importance"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lhZTPyZLuK-J",
        "outputId": "8be4ae16-cb54-41ad-9a3b-25599d724497"
      },
      "source": [
        "estimator = RandomForestClassifier(n_estimators=1000, min_samples_split=6,\n",
        "                                    criterion='gini', max_features='auto', oob_score=True,random_state=1,n_jobs=-1)\n",
        "estimator.fit(X_train, y_train)\n",
        "\n",
        "training_accuracy = estimator.score(X_train, y_train)   \n",
        "print('Training accuracy: ', training_accuracy)                          \n",
        "validation_accuracy = estimator.score(X_val, y_val)\n",
        "print('Validation accuracy: ', validation_accuracy)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training accuracy:  0.8970819304152637\n",
            "Validation accuracy:  0.7950617283950617\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j9_cgCPFjyrz",
        "outputId": "73461ee4-ac53-45fb-a2e8-a46785e14bbe"
      },
      "source": [
        "print(estimator.feature_importances_)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[3.47055073e-02 1.42967611e-01 2.53851206e-03 7.00456024e-02\n",
            " 3.49063449e-03 1.00223983e-02 3.33187094e-03 1.23686725e-02\n",
            " 2.25989387e-03 3.09421703e-03 1.47579681e-03 1.09861254e-03\n",
            " 1.25910642e-02 4.14329957e-03 1.16705396e-02 3.90600690e-03\n",
            " 9.48857234e-03 5.98554642e-03 5.28654789e-03 8.37169621e-03\n",
            " 9.12959533e-03 7.46239248e-03 6.96953378e-03 6.44732317e-03\n",
            " 6.66653798e-03 7.44153374e-03 8.61178502e-03 4.11375247e-03\n",
            " 9.86667338e-03 1.23237480e-02 5.44035120e-03 3.75230611e-03\n",
            " 3.79006806e-03 6.36384635e-03 1.10670305e-02 1.13703437e-02\n",
            " 4.26301517e-03 1.23060255e-02 1.05085252e-02 1.63929339e-03\n",
            " 1.01508587e-02 8.62322393e-03 8.05799941e-03 8.41103445e-03\n",
            " 1.16757380e-02 8.13944022e-03 4.43587897e-03 2.77140891e-02\n",
            " 1.31701798e-03 6.57737233e-03 3.51628330e-04 4.58368281e-03\n",
            " 7.90945410e-03 1.71802344e-02 4.78892281e-03 1.65837254e-03\n",
            " 8.46866690e-03 8.57858238e-03 1.59680699e-03 7.18465197e-04\n",
            " 2.07224969e-05 1.82544402e-03 5.56440330e-03 1.49392296e-03\n",
            " 9.03825772e-03 7.65883039e-03 1.03165330e-01 3.61137490e-02\n",
            " 1.95745552e-02 9.10249867e-03 2.62065624e-03 8.73615788e-03\n",
            " 2.11914334e-03 6.35984199e-04 4.68208022e-03 5.66767169e-03\n",
            " 7.70519577e-03 9.05497319e-03 6.09955184e-03 6.10437679e-03\n",
            " 6.36631407e-04 2.60162277e-04 1.23797977e-02 1.12374236e-02\n",
            " 2.85941108e-05 5.38467148e-03 1.62112275e-03 2.88441913e-02\n",
            " 2.46969603e-04 5.42452391e-03 3.24260969e-05 6.10827736e-03\n",
            " 1.53162619e-03 2.99663254e-02]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "HUsKRUEMj_Zt",
        "outputId": "ad3cd7e4-0ac1-4f61-819e-9d9560a9f16d"
      },
      "source": [
        "importance = estimator.feature_importances_\n",
        "importance = pd.DataFrame(importance, index=X_train.columns, columns=[\"Importance\"])\n",
        "\n",
        "importance[\"Std\"] = np.std([tree.feature_importances_ for tree in estimator.estimators_], axis=0)\n",
        "\n",
        "x = range(importance.shape[0])\n",
        "y = importance.iloc[:, 0]\n",
        "yerr = importance.iloc[:, 1]\n",
        "lables = list(train_df.columns.values)\n",
        "\n",
        "\n",
        "plt.figure(figsize=(20,20))\n",
        "ax = plt.bar(x, y, yerr=yerr ,align=\"center\")\n",
        "plt.xlabel('Features', fontsize = 12)\n",
        "plt.ylabel('Importance', fontsize = 12)\n",
        "\n",
        "#plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Importance')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 771
        },
        "id": "paGCTuNblq6F",
        "outputId": "dfc40047-a2d8-4fca-e2e0-e3e6c5994366"
      },
      "source": [
        "pd.set_option('display.max_rows', 100)\n",
        "major_features=importance[importance.Importance>=0.01]\n",
        "display(major_features)\n",
        "pd.set_option('display.max_rows', 15)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Importance</th>\n",
              "      <th>Std</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>amount_tsh</th>\n",
              "      <td>0.034706</td>\n",
              "      <td>0.013641</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>days_since_recorded</th>\n",
              "      <td>0.142968</td>\n",
              "      <td>0.006037</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>population</th>\n",
              "      <td>0.070046</td>\n",
              "      <td>0.004568</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>funder_gov</th>\n",
              "      <td>0.010022</td>\n",
              "      <td>0.003371</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>funder_other</th>\n",
              "      <td>0.012369</td>\n",
              "      <td>0.003148</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>installer_dwe</th>\n",
              "      <td>0.012591</td>\n",
              "      <td>0.002344</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>installer_other</th>\n",
              "      <td>0.011671</td>\n",
              "      <td>0.002050</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>scheme_management_vwc</th>\n",
              "      <td>0.012324</td>\n",
              "      <td>0.002280</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>permit_False</th>\n",
              "      <td>0.011067</td>\n",
              "      <td>0.002043</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>permit_True</th>\n",
              "      <td>0.011370</td>\n",
              "      <td>0.002250</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>construction_year_00s</th>\n",
              "      <td>0.012306</td>\n",
              "      <td>0.003094</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>construction_year_10s</th>\n",
              "      <td>0.010509</td>\n",
              "      <td>0.002524</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>construction_year_70s</th>\n",
              "      <td>0.010151</td>\n",
              "      <td>0.003879</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>extraction_type_class_gravity</th>\n",
              "      <td>0.011676</td>\n",
              "      <td>0.007583</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>extraction_type_class_other</th>\n",
              "      <td>0.027714</td>\n",
              "      <td>0.031314</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>payment_type_never pay</th>\n",
              "      <td>0.017180</td>\n",
              "      <td>0.006591</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>quantity_dry</th>\n",
              "      <td>0.103165</td>\n",
              "      <td>0.044952</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>quantity_enough</th>\n",
              "      <td>0.036114</td>\n",
              "      <td>0.018203</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>quantity_insufficient</th>\n",
              "      <td>0.019575</td>\n",
              "      <td>0.011440</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>waterpoint_type_communal standpipe</th>\n",
              "      <td>0.012380</td>\n",
              "      <td>0.010041</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>waterpoint_type_communal standpipe multiple</th>\n",
              "      <td>0.011237</td>\n",
              "      <td>0.005700</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>waterpoint_type_other</th>\n",
              "      <td>0.028844</td>\n",
              "      <td>0.036870</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>waterpoint_type_group_other</th>\n",
              "      <td>0.029966</td>\n",
              "      <td>0.037306</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                             Importance       Std\n",
              "amount_tsh                                     0.034706  0.013641\n",
              "days_since_recorded                            0.142968  0.006037\n",
              "population                                     0.070046  0.004568\n",
              "funder_gov                                     0.010022  0.003371\n",
              "funder_other                                   0.012369  0.003148\n",
              "installer_dwe                                  0.012591  0.002344\n",
              "installer_other                                0.011671  0.002050\n",
              "scheme_management_vwc                          0.012324  0.002280\n",
              "permit_False                                   0.011067  0.002043\n",
              "permit_True                                    0.011370  0.002250\n",
              "construction_year_00s                          0.012306  0.003094\n",
              "construction_year_10s                          0.010509  0.002524\n",
              "construction_year_70s                          0.010151  0.003879\n",
              "extraction_type_class_gravity                  0.011676  0.007583\n",
              "extraction_type_class_other                    0.027714  0.031314\n",
              "payment_type_never pay                         0.017180  0.006591\n",
              "quantity_dry                                   0.103165  0.044952\n",
              "quantity_enough                                0.036114  0.018203\n",
              "quantity_insufficient                          0.019575  0.011440\n",
              "waterpoint_type_communal standpipe             0.012380  0.010041\n",
              "waterpoint_type_communal standpipe multiple    0.011237  0.005700\n",
              "waterpoint_type_other                          0.028844  0.036870\n",
              "waterpoint_type_group_other                    0.029966  0.037306"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pG2W4sZ7luTn",
        "outputId": "9ad18526-6fab-4426-8f3a-d6e775191331"
      },
      "source": [
        "selected_features = list(major_features.index.values)\n",
        "selected_features"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['amount_tsh',\n",
              " 'days_since_recorded',\n",
              " 'population',\n",
              " 'funder_gov',\n",
              " 'funder_other',\n",
              " 'installer_dwe',\n",
              " 'installer_other',\n",
              " 'scheme_management_vwc',\n",
              " 'permit_False',\n",
              " 'permit_True',\n",
              " 'construction_year_00s',\n",
              " 'construction_year_10s',\n",
              " 'construction_year_70s',\n",
              " 'extraction_type_class_gravity',\n",
              " 'extraction_type_class_other',\n",
              " 'payment_type_never pay',\n",
              " 'quantity_dry',\n",
              " 'quantity_enough',\n",
              " 'quantity_insufficient',\n",
              " 'waterpoint_type_communal standpipe',\n",
              " 'waterpoint_type_communal standpipe multiple',\n",
              " 'waterpoint_type_other',\n",
              " 'waterpoint_type_group_other']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LuwgF-BasaPP"
      },
      "source": [
        "new_X_train = X_train[selected_features]\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JzF3TRmUtGVr"
      },
      "source": [
        "new_X_val = X_val[selected_features]\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ia11nMb9tjfc",
        "outputId": "566cfbd7-6e05-4ff5-dd03-8f55e5c40682"
      },
      "source": [
        "estimator = RandomForestClassifier(n_estimators=1000, min_samples_split=6,\n",
        "                                    criterion='gini', max_features='auto', oob_score=True,random_state=1,n_jobs=-1)\n",
        "estimator.fit(new_X_train, y_train)\n",
        "\n",
        "training_accuracy = estimator.score(new_X_train, y_train)   \n",
        "print('Training accuracy: ', training_accuracy)                          \n",
        "validation_accuracy = estimator.score(new_X_val, y_val)\n",
        "print('Validation accuracy: ', validation_accuracy)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training accuracy:  0.8874859708193041\n",
            "Validation accuracy:  0.7800224466891134\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SVzbfgouuaVf"
      },
      "source": [
        "**Using complete set of features:**\n",
        "\n",
        "    Training accuracy:  0.8970819304152637\n",
        "    Validation accuracy:  0.7950617283950617\n",
        "\n",
        "**Using Selected features based on importnace**\n",
        "\n",
        "    Training accuracy:  0.8874859708193041\n",
        "    Validation accuracy:  0.7800224466891134\n",
        "\n",
        "We can see that we are getting almost similar accuracy for our model with lesser features there by making the model computationally less expensive"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FolEvY8jxh5F"
      },
      "source": [
        "Also we can also identify that \"quantity\" and \"day_since_Recorded\" are the two msot important features"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wia6a2Zqo3i7"
      },
      "source": [
        "# Visualising Boundaries"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6LExK8CeCF17"
      },
      "source": [
        "Visualising Decision Tree"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CBH8lt07AgRz",
        "outputId": "e3afde38-d2f1-4ad4-fb2c-99a1083a9b7a"
      },
      "source": [
        "\n",
        "decisiontree = DecisionTreeClassifier()\n",
        "model = decisiontree.fit(X_train,y_train)\n",
        "\n",
        "dot_data = tree.export_graphviz(decisiontree,out_file = None,feature_names = features,\n",
        "                            class_names=classes)\n",
        "graph = pydotplus.graph_from_dot_data(dot_data)\n",
        "\n",
        "graph.write_png('tree.png')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "dot: graph is too large for cairo-renderer bitmaps. Scaling by 0.0713227 to fit\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 98
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t4kUfPlSDLAc"
      },
      "source": [
        "Attaching 'tree.png' along with the files uploaded"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RHZiEEcpTWTB"
      },
      "source": [
        "# References:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sLiEmrJ0W7rS"
      },
      "source": [
        "\n",
        "\n",
        "1.   https://towardsdatascience.com/predicting-the-functional-status-of-pumps-in-tanzania-355c9269d0c2\n",
        "2.   https://towardsdatascience.com/water-pumps-maintenance-prediction-data-science-illustrated-20c7100017c5\n",
        "3.  https://www.geeksforgeeks.org/python-pandas-get_dummies-method/\n",
        "4. https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.get_dummies.html\n",
        "5.  https://www.marsja.se/how-to-use-pandas-get_dummies-to-create-dummy-variables-in-python/\n",
        "6. https://koalatea.io/sklearn-decision-tree-vizualization/\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QK97qf4ZXCtg"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}