5-feature-selection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "5d56f523",
   "metadata": {},
   "source": [
    "# First Floor Height Feature Selection\n",
    "\n",
    "Exploring the NYC Building Elevation and Subgrade (BES) Data Set in Python  \n",
    "\n",
    "Author: Mark Bauer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "55b4ccdd",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import glob\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib as mpl\n",
    "import seaborn as sns\n",
    "import duckdb\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.inspection import permutation_importance\n",
    "\n",
    "plt.rcParams['savefig.facecolor'] = 'white'\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c14469e6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Python implementation: CPython\n",
      "Python version       : 3.8.13\n",
      "IPython version      : 8.4.0\n",
      "\n",
      "pandas    : 1.4.3\n",
      "matplotlib: 3.5.2\n",
      "seaborn   : 0.11.2\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Printing verions of Python modules and packages with **watermark** - the IPython magic extension.\n",
    "%reload_ext watermark\n",
    "%watermark -v -p pandas,matplotlib,seaborn"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c7b16a1",
   "metadata": {},
   "source": [
    "# Dataset: Building Elevation and Subgrade (BES)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b989c0f6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "┌────────────┐\n",
      "│ count_rows │\n",
      "│   int64    │\n",
      "├────────────┤\n",
      "│     861876 │\n",
      "└────────────┘\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# create a DuckDB database instance\n",
    "con = duckdb.connect()\n",
    "\n",
    "# examine count of rows\n",
    "con.sql(\"SELECT COUNT(*) AS count_rows FROM read_parquet('data/bes-data.parquet')\").show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "99389ded",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rows: 861,876\n",
      "columns: 26\n"
     ]
    },
    {
     "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>the_geom</th>\n",
       "      <th>bin</th>\n",
       "      <th>bbl</th>\n",
       "      <th>borough</th>\n",
       "      <th>block</th>\n",
       "      <th>lot</th>\n",
       "      <th>address</th>\n",
       "      <th>z_grade</th>\n",
       "      <th>z_floor</th>\n",
       "      <th>subgrade</th>\n",
       "      <th>...</th>\n",
       "      <th>longitude</th>\n",
       "      <th>pluto_bbl</th>\n",
       "      <th>Council</th>\n",
       "      <th>BoroCD</th>\n",
       "      <th>CTLabel</th>\n",
       "      <th>BoroCT2020</th>\n",
       "      <th>NTA2020</th>\n",
       "      <th>NTAName</th>\n",
       "      <th>CDTA2020</th>\n",
       "      <th>CDTAName</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>POINT (-74.22274561187417 40.52134422844183)</td>\n",
       "      <td>5128004</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7534.0</td>\n",
       "      <td>353</td>\n",
       "      <td>78 SAVO LOOP</td>\n",
       "      <td>29.428</td>\n",
       "      <td>32.332</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.222746</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>226.01</td>\n",
       "      <td>5022601</td>\n",
       "      <td>SI0304</td>\n",
       "      <td>Annadale-Huguenot-Prince's Bay-Woodrow</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>POINT (-74.24179250549321 40.52875316810818)</td>\n",
       "      <td>5155392</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7596.0</td>\n",
       "      <td>125</td>\n",
       "      <td>72 CHART LOOP</td>\n",
       "      <td>25.366</td>\n",
       "      <td>26.703</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.241793</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>226.02</td>\n",
       "      <td>5022602</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>POINT (-74.24109012652094 40.528883822921635)</td>\n",
       "      <td>5148808</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7596.0</td>\n",
       "      <td>131</td>\n",
       "      <td>40 TIDES LANE</td>\n",
       "      <td>36.172</td>\n",
       "      <td>38.902</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.241090</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>226.02</td>\n",
       "      <td>5022602</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>POINT (-74.24649640323727 40.507045708438184)</td>\n",
       "      <td>5088274</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7915.0</td>\n",
       "      <td>42</td>\n",
       "      <td>328 MAIN STREET</td>\n",
       "      <td>69.897</td>\n",
       "      <td>72.459</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.246496</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>244.01</td>\n",
       "      <td>5024401</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>POINT (-74.24198032271697 40.51005666426356)</td>\n",
       "      <td>5087850</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7868.0</td>\n",
       "      <td>123</td>\n",
       "      <td>309 SLEIGHT AVENUE</td>\n",
       "      <td>74.907</td>\n",
       "      <td>79.122</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.241980</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>244.01</td>\n",
       "      <td>5024401</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                        the_geom      bin         bbl  \\\n",
       "0   POINT (-74.22274561187417 40.52134422844183)  5128004  5075340353   \n",
       "1   POINT (-74.24179250549321 40.52875316810818)  5155392  5075960125   \n",
       "2  POINT (-74.24109012652094 40.528883822921635)  5148808  5075960131   \n",
       "3  POINT (-74.24649640323727 40.507045708438184)  5088274  5079150042   \n",
       "4   POINT (-74.24198032271697 40.51005666426356)  5087850  5078680123   \n",
       "\n",
       "   borough   block  lot             address  z_grade  z_floor subgrade  ...  \\\n",
       "0      5.0  7534.0  353        78 SAVO LOOP   29.428   32.332        N  ...   \n",
       "1      5.0  7596.0  125       72 CHART LOOP   25.366   26.703        N  ...   \n",
       "2      5.0  7596.0  131       40 TIDES LANE   36.172   38.902        N  ...   \n",
       "3      5.0  7915.0   42     328 MAIN STREET   69.897   72.459        N  ...   \n",
       "4      5.0  7868.0  123  309 SLEIGHT AVENUE   74.907   79.122        N  ...   \n",
       "\n",
       "   longitude   pluto_bbl Council  BoroCD  CTLabel  BoroCT2020  NTA2020  \\\n",
       "0 -74.222746  5075340353      51     503   226.01     5022601   SI0304   \n",
       "1 -74.241793  5075960125      51     503   226.02     5022602   SI0305   \n",
       "2 -74.241090  5075960131      51     503   226.02     5022602   SI0305   \n",
       "3 -74.246496  5079150042      51     503   244.01     5024401   SI0305   \n",
       "4 -74.241980  5078680123      51     503   244.01     5024401   SI0305   \n",
       "\n",
       "                                  NTAName  CDTA2020  \\\n",
       "0  Annadale-Huguenot-Prince's Bay-Woodrow      SI03   \n",
       "1                  Tottenville-Charleston      SI03   \n",
       "2                  Tottenville-Charleston      SI03   \n",
       "3                  Tottenville-Charleston      SI03   \n",
       "4                  Tottenville-Charleston      SI03   \n",
       "\n",
       "                                CDTAName  \n",
       "0  SI03 South Shore (CD 3 Approximation)  \n",
       "1  SI03 South Shore (CD 3 Approximation)  \n",
       "2  SI03 South Shore (CD 3 Approximation)  \n",
       "3  SI03 South Shore (CD 3 Approximation)  \n",
       "4  SI03 South Shore (CD 3 Approximation)  \n",
       "\n",
       "[5 rows x 26 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bes_df = con.sql(\"SELECT * FROM read_parquet('data/bes-data.parquet')\").df()\n",
    "\n",
    "rows, cols = bes_df.shape\n",
    "print('rows: {:,}\\ncolumns: {}'.format(rows, cols))\n",
    "\n",
    "# preview data\n",
    "bes_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c73c5e31",
   "metadata": {},
   "source": [
    "`z_grade`: The elevation of the building at it's lowest adjacent grade - the lowest point where the building touches the ground.\n",
    "\n",
    "`z_floor`: The elevation of what is estimated to be the lowest actively used floor.\n",
    "\n",
    "Definitions from the data dictionary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "0079748c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "new shape of data: (646781, 26)\n"
     ]
    },
    {
     "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>the_geom</th>\n",
       "      <th>bin</th>\n",
       "      <th>bbl</th>\n",
       "      <th>borough</th>\n",
       "      <th>block</th>\n",
       "      <th>lot</th>\n",
       "      <th>address</th>\n",
       "      <th>z_grade</th>\n",
       "      <th>z_floor</th>\n",
       "      <th>subgrade</th>\n",
       "      <th>...</th>\n",
       "      <th>longitude</th>\n",
       "      <th>pluto_bbl</th>\n",
       "      <th>Council</th>\n",
       "      <th>BoroCD</th>\n",
       "      <th>CTLabel</th>\n",
       "      <th>BoroCT2020</th>\n",
       "      <th>NTA2020</th>\n",
       "      <th>NTAName</th>\n",
       "      <th>CDTA2020</th>\n",
       "      <th>CDTAName</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>POINT (-74.22274561187417 40.52134422844183)</td>\n",
       "      <td>5128004</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7534.0</td>\n",
       "      <td>353</td>\n",
       "      <td>78 SAVO LOOP</td>\n",
       "      <td>29.428</td>\n",
       "      <td>32.332</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.222746</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>226.01</td>\n",
       "      <td>5022601</td>\n",
       "      <td>SI0304</td>\n",
       "      <td>Annadale-Huguenot-Prince's Bay-Woodrow</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>POINT (-74.24179250549321 40.52875316810818)</td>\n",
       "      <td>5155392</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7596.0</td>\n",
       "      <td>125</td>\n",
       "      <td>72 CHART LOOP</td>\n",
       "      <td>25.366</td>\n",
       "      <td>26.703</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.241793</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>226.02</td>\n",
       "      <td>5022602</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>POINT (-74.24109012652094 40.528883822921635)</td>\n",
       "      <td>5148808</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7596.0</td>\n",
       "      <td>131</td>\n",
       "      <td>40 TIDES LANE</td>\n",
       "      <td>36.172</td>\n",
       "      <td>38.902</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.241090</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>226.02</td>\n",
       "      <td>5022602</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>POINT (-74.24649640323727 40.507045708438184)</td>\n",
       "      <td>5088274</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7915.0</td>\n",
       "      <td>42</td>\n",
       "      <td>328 MAIN STREET</td>\n",
       "      <td>69.897</td>\n",
       "      <td>72.459</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.246496</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>244.01</td>\n",
       "      <td>5024401</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>POINT (-74.24198032271697 40.51005666426356)</td>\n",
       "      <td>5087850</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7868.0</td>\n",
       "      <td>123</td>\n",
       "      <td>309 SLEIGHT AVENUE</td>\n",
       "      <td>74.907</td>\n",
       "      <td>79.122</td>\n",
       "      <td>N</td>\n",
       "      <td>...</td>\n",
       "      <td>-74.241980</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>51</td>\n",
       "      <td>503</td>\n",
       "      <td>244.01</td>\n",
       "      <td>5024401</td>\n",
       "      <td>SI0305</td>\n",
       "      <td>Tottenville-Charleston</td>\n",
       "      <td>SI03</td>\n",
       "      <td>SI03 South Shore (CD 3 Approximation)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                        the_geom      bin         bbl  \\\n",
       "0   POINT (-74.22274561187417 40.52134422844183)  5128004  5075340353   \n",
       "1   POINT (-74.24179250549321 40.52875316810818)  5155392  5075960125   \n",
       "2  POINT (-74.24109012652094 40.528883822921635)  5148808  5075960131   \n",
       "3  POINT (-74.24649640323727 40.507045708438184)  5088274  5079150042   \n",
       "4   POINT (-74.24198032271697 40.51005666426356)  5087850  5078680123   \n",
       "\n",
       "   borough   block  lot             address  z_grade  z_floor subgrade  ...  \\\n",
       "0      5.0  7534.0  353        78 SAVO LOOP   29.428   32.332        N  ...   \n",
       "1      5.0  7596.0  125       72 CHART LOOP   25.366   26.703        N  ...   \n",
       "2      5.0  7596.0  131       40 TIDES LANE   36.172   38.902        N  ...   \n",
       "3      5.0  7915.0   42     328 MAIN STREET   69.897   72.459        N  ...   \n",
       "4      5.0  7868.0  123  309 SLEIGHT AVENUE   74.907   79.122        N  ...   \n",
       "\n",
       "   longitude   pluto_bbl Council  BoroCD  CTLabel  BoroCT2020  NTA2020  \\\n",
       "0 -74.222746  5075340353      51     503   226.01     5022601   SI0304   \n",
       "1 -74.241793  5075960125      51     503   226.02     5022602   SI0305   \n",
       "2 -74.241090  5075960131      51     503   226.02     5022602   SI0305   \n",
       "3 -74.246496  5079150042      51     503   244.01     5024401   SI0305   \n",
       "4 -74.241980  5078680123      51     503   244.01     5024401   SI0305   \n",
       "\n",
       "                                  NTAName  CDTA2020  \\\n",
       "0  Annadale-Huguenot-Prince's Bay-Woodrow      SI03   \n",
       "1                  Tottenville-Charleston      SI03   \n",
       "2                  Tottenville-Charleston      SI03   \n",
       "3                  Tottenville-Charleston      SI03   \n",
       "4                  Tottenville-Charleston      SI03   \n",
       "\n",
       "                                CDTAName  \n",
       "0  SI03 South Shore (CD 3 Approximation)  \n",
       "1  SI03 South Shore (CD 3 Approximation)  \n",
       "2  SI03 South Shore (CD 3 Approximation)  \n",
       "3  SI03 South Shore (CD 3 Approximation)  \n",
       "4  SI03 South Shore (CD 3 Approximation)  \n",
       "\n",
       "[5 rows x 26 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bes_df = (\n",
    "    bes_df\n",
    "    .loc[\n",
    "        (bes_df['z_floor'] != 0)\n",
    "        & (bes_df['z_grade'] != 0)\n",
    "        & (bes_df['notes1'] == 'Property was Successfully Measured')]\n",
    "    .reset_index(drop=True)\n",
    ")\n",
    "\n",
    "print('new shape of data:', bes_df.shape)\n",
    "bes_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee0c05dc",
   "metadata": {},
   "source": [
    "Calculating Height from Building Adjacent Grade to First Floor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "a6dc8873",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    646781.000000\n",
       "mean          3.834224\n",
       "std           2.880029\n",
       "min          -8.832000\n",
       "25%           1.671000\n",
       "50%           3.365000\n",
       "75%           5.009000\n",
       "max          83.759000\n",
       "Name: z_floor_grade, dtype: float64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# height from building floor to grade\n",
    "bes_df['z_floor_grade'] = bes_df['z_floor'] - bes_df['z_grade']\n",
    "\n",
    "# summary stats\n",
    "bes_df['z_floor_grade'].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "387dec51",
   "metadata": {},
   "source": [
    "# Dataset: NYC Building Footprints"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "96266faa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rows: 1,083,074\n",
      "columns: 16\n"
     ]
    },
    {
     "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>the_geom</th>\n",
       "      <th>NAME</th>\n",
       "      <th>BIN</th>\n",
       "      <th>CNSTRCT_YR</th>\n",
       "      <th>LSTMODDATE</th>\n",
       "      <th>LSTSTATYPE</th>\n",
       "      <th>DOITT_ID</th>\n",
       "      <th>HEIGHTROOF</th>\n",
       "      <th>FEAT_CODE</th>\n",
       "      <th>GROUNDELEV</th>\n",
       "      <th>SHAPE_AREA</th>\n",
       "      <th>SHAPE_LEN</th>\n",
       "      <th>BASE_BBL</th>\n",
       "      <th>MPLUTO_BBL</th>\n",
       "      <th>GEOMSOURCE</th>\n",
       "      <th>GLOBALID</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>MULTIPOLYGON (((-73.96664570466969 40.62599676...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3170958</td>\n",
       "      <td>1925.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>96807</td>\n",
       "      <td>29.749853</td>\n",
       "      <td>2100</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3065220021</td>\n",
       "      <td>3065220021</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>{31298F86-3088-4F53-B3DB-71A9EFA6FA1F}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MULTIPOLYGON (((-74.16790202462265 40.63936048...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5028452</td>\n",
       "      <td>1965.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>326368</td>\n",
       "      <td>22.630000</td>\n",
       "      <td>2100</td>\n",
       "      <td>39.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5012640036</td>\n",
       "      <td>5012640036</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>{F5F8CDA5-69E2-46F8-8F69-BA95C025B520}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MULTIPOLYGON (((-74.19510813278613 40.55610681...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5078368</td>\n",
       "      <td>1970.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>746627</td>\n",
       "      <td>35.760000</td>\n",
       "      <td>2100</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5060190091</td>\n",
       "      <td>5060190091</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>{9F644794-F72C-4582-9E5E-B337E2B97068}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MULTIPOLYGON (((-73.96113466505085 40.57743931...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3245111</td>\n",
       "      <td>1928.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>786626</td>\n",
       "      <td>37.500000</td>\n",
       "      <td>2100</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3086910048</td>\n",
       "      <td>3086910048</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>{F916B22D-E25B-44AE-9FA9-2A51191B9CDF}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>MULTIPOLYGON (((-73.75421559146166 40.75591276...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4161096</td>\n",
       "      <td>1950.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>746409</td>\n",
       "      <td>18.015113</td>\n",
       "      <td>2100</td>\n",
       "      <td>93.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4075020005</td>\n",
       "      <td>4075020005</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>{525F2C24-616B-4F29-98A3-8FEA5D4B1A7D}</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            the_geom NAME      BIN  \\\n",
       "0  MULTIPOLYGON (((-73.96664570466969 40.62599676...  NaN  3170958   \n",
       "1  MULTIPOLYGON (((-74.16790202462265 40.63936048...  NaN  5028452   \n",
       "2  MULTIPOLYGON (((-74.19510813278613 40.55610681...  NaN  5078368   \n",
       "3  MULTIPOLYGON (((-73.96113466505085 40.57743931...  NaN  3245111   \n",
       "4  MULTIPOLYGON (((-73.75421559146166 40.75591276...  NaN  4161096   \n",
       "\n",
       "   CNSTRCT_YR              LSTMODDATE   LSTSTATYPE  DOITT_ID  HEIGHTROOF  \\\n",
       "0      1925.0  08/22/2017 12:00:00 AM  Constructed     96807   29.749853   \n",
       "1      1965.0  08/22/2017 12:00:00 AM  Constructed    326368   22.630000   \n",
       "2      1970.0  08/22/2017 12:00:00 AM  Constructed    746627   35.760000   \n",
       "3      1928.0  08/22/2017 12:00:00 AM  Constructed    786626   37.500000   \n",
       "4      1950.0  08/22/2017 12:00:00 AM  Constructed    746409   18.015113   \n",
       "\n",
       "   FEAT_CODE  GROUNDELEV  SHAPE_AREA  SHAPE_LEN    BASE_BBL  MPLUTO_BBL  \\\n",
       "0       2100        40.0           0          0  3065220021  3065220021   \n",
       "1       2100        39.0           0          0  5012640036  5012640036   \n",
       "2       2100        51.0           0          0  5060190091  5060190091   \n",
       "3       2100         6.0           0          0  3086910048  3086910048   \n",
       "4       2100        93.0           0          0  4075020005  4075020005   \n",
       "\n",
       "   GEOMSOURCE                                GLOBALID  \n",
       "0  Photogramm  {31298F86-3088-4F53-B3DB-71A9EFA6FA1F}  \n",
       "1  Photogramm  {F5F8CDA5-69E2-46F8-8F69-BA95C025B520}  \n",
       "2  Photogramm  {9F644794-F72C-4582-9E5E-B337E2B97068}  \n",
       "3  Photogramm  {F916B22D-E25B-44AE-9FA9-2A51191B9CDF}  \n",
       "4  Photogramm  {525F2C24-616B-4F29-98A3-8FEA5D4B1A7D}  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = \"https://data.cityofnewyork.us/api/views/qb5r-6dgf/rows.csv?accessType=DOWNLOAD\"\n",
    "buildings_df = pd.read_csv(path, low_memory=False)\n",
    "\n",
    "rows, cols = buildings_df.shape\n",
    "print('rows: {:,}\\ncolumns: {}'.format(rows, cols))\n",
    "\n",
    "# preview data\n",
    "buildings_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f78fd97c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1083074 entries, 0 to 1083073\n",
      "Data columns (total 16 columns):\n",
      " #   Column      Non-Null Count    Dtype  \n",
      "---  ------      --------------    -----  \n",
      " 0   the_geom    1083074 non-null  object \n",
      " 1   NAME        2254 non-null     object \n",
      " 2   BIN         1083074 non-null  int64  \n",
      " 3   CNSTRCT_YR  1072806 non-null  float64\n",
      " 4   LSTMODDATE  1083074 non-null  object \n",
      " 5   LSTSTATYPE  1082746 non-null  object \n",
      " 6   DOITT_ID    1083074 non-null  int64  \n",
      " 7   HEIGHTROOF  1080494 non-null  float64\n",
      " 8   FEAT_CODE   1083074 non-null  int64  \n",
      " 9   GROUNDELEV  1082522 non-null  float64\n",
      " 10  SHAPE_AREA  1083074 non-null  int64  \n",
      " 11  SHAPE_LEN   1083074 non-null  int64  \n",
      " 12  BASE_BBL    1083074 non-null  int64  \n",
      " 13  MPLUTO_BBL  1083074 non-null  int64  \n",
      " 14  GEOMSOURCE  1082778 non-null  object \n",
      " 15  GLOBALID    1083074 non-null  object \n",
      "dtypes: float64(3), int64(7), object(6)\n",
      "memory usage: 132.2+ MB\n"
     ]
    }
   ],
   "source": [
    "buildings_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "479ec096",
   "metadata": {},
   "source": [
    "# Dataset: NYC PLUTO"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "5c0fadc6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rows: 858,572\n",
      "columns: 101\n"
     ]
    },
    {
     "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>borough</th>\n",
       "      <th>block</th>\n",
       "      <th>lot</th>\n",
       "      <th>community board</th>\n",
       "      <th>census tract 2010</th>\n",
       "      <th>cb2010</th>\n",
       "      <th>schooldist</th>\n",
       "      <th>council district</th>\n",
       "      <th>postcode</th>\n",
       "      <th>firecomp</th>\n",
       "      <th>...</th>\n",
       "      <th>landmkdate</th>\n",
       "      <th>basempdate</th>\n",
       "      <th>masdate</th>\n",
       "      <th>polidate</th>\n",
       "      <th>edesigdate</th>\n",
       "      <th>geom</th>\n",
       "      <th>dcpedited</th>\n",
       "      <th>notes</th>\n",
       "      <th>bct2020</th>\n",
       "      <th>bctcb2020</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>QN</td>\n",
       "      <td>798</td>\n",
       "      <td>31</td>\n",
       "      <td>401.0</td>\n",
       "      <td>123.01</td>\n",
       "      <td>2006.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>11105.0</td>\n",
       "      <td>E263</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4012301.0</td>\n",
       "      <td>4.012301e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BK</td>\n",
       "      <td>6429</td>\n",
       "      <td>16</td>\n",
       "      <td>311.0</td>\n",
       "      <td>172.00</td>\n",
       "      <td>2004.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>11228.0</td>\n",
       "      <td>L149</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3017200.0</td>\n",
       "      <td>3.017200e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BK</td>\n",
       "      <td>7790</td>\n",
       "      <td>42</td>\n",
       "      <td>318.0</td>\n",
       "      <td>732.00</td>\n",
       "      <td>1004.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>11210.0</td>\n",
       "      <td>L159</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3073200.0</td>\n",
       "      <td>3.073200e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BK</td>\n",
       "      <td>7197</td>\n",
       "      <td>52</td>\n",
       "      <td>315.0</td>\n",
       "      <td>374.01</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>11223.0</td>\n",
       "      <td>E245</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3037401.0</td>\n",
       "      <td>3.037401e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>QN</td>\n",
       "      <td>15318</td>\n",
       "      <td>35</td>\n",
       "      <td>414.0</td>\n",
       "      <td>1072.01</td>\n",
       "      <td>1032.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>11693.0</td>\n",
       "      <td>E266</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4107201.0</td>\n",
       "      <td>4.107201e+10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 101 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  borough  block  lot  community board  census tract 2010  cb2010  schooldist  \\\n",
       "0      QN    798   31            401.0             123.01  2006.0        30.0   \n",
       "1      BK   6429   16            311.0             172.00  2004.0        20.0   \n",
       "2      BK   7790   42            318.0             732.00  1004.0        22.0   \n",
       "3      BK   7197   52            315.0             374.01  2000.0        21.0   \n",
       "4      QN  15318   35            414.0            1072.01  1032.0        27.0   \n",
       "\n",
       "   council district  postcode firecomp  ...  landmkdate  basempdate  masdate  \\\n",
       "0              22.0   11105.0     E263  ...         NaN         NaN      NaN   \n",
       "1              50.0   11228.0     L149  ...         NaN         NaN      NaN   \n",
       "2              45.0   11210.0     L159  ...         NaN         NaN      NaN   \n",
       "3              47.0   11223.0     E245  ...         NaN         NaN      NaN   \n",
       "4              32.0   11693.0     E266  ...         NaN         NaN      NaN   \n",
       "\n",
       "  polidate edesigdate geom dcpedited notes    bct2020     bctcb2020  \n",
       "0      NaN        NaN  NaN       NaN   NaN  4012301.0  4.012301e+10  \n",
       "1      NaN        NaN  NaN       NaN   NaN  3017200.0  3.017200e+10  \n",
       "2      NaN        NaN  NaN       NaN   NaN  3073200.0  3.073200e+10  \n",
       "3      NaN        NaN  NaN       NaN   NaN  3037401.0  3.037401e+10  \n",
       "4      NaN        NaN  NaN       NaN   NaN  4107201.0  4.107201e+10  \n",
       "\n",
       "[5 rows x 101 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = \"https://data.cityofnewyork.us/api/views/64uk-42ks/rows.csv?accessType=DOWNLOAD\"\n",
    "pluto_df = pd.read_csv(path, low_memory=False)\n",
    "\n",
    "rows, cols = pluto_df.shape\n",
    "print('rows: {:,}\\ncolumns: {}'.format(rows, cols))\n",
    "\n",
    "# preview data\n",
    "pluto_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "de75da38",
   "metadata": {},
   "outputs": [
    {
     "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>borough</th>\n",
       "      <th>block</th>\n",
       "      <th>lot</th>\n",
       "      <th>community board</th>\n",
       "      <th>census tract 2010</th>\n",
       "      <th>cb2010</th>\n",
       "      <th>schooldist</th>\n",
       "      <th>council district</th>\n",
       "      <th>postcode</th>\n",
       "      <th>firecomp</th>\n",
       "      <th>...</th>\n",
       "      <th>landmkdate</th>\n",
       "      <th>basempdate</th>\n",
       "      <th>masdate</th>\n",
       "      <th>polidate</th>\n",
       "      <th>edesigdate</th>\n",
       "      <th>geom</th>\n",
       "      <th>dcpedited</th>\n",
       "      <th>notes</th>\n",
       "      <th>bct2020</th>\n",
       "      <th>bctcb2020</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>QN</td>\n",
       "      <td>798</td>\n",
       "      <td>31</td>\n",
       "      <td>401.0</td>\n",
       "      <td>123.01</td>\n",
       "      <td>2006.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>11105.0</td>\n",
       "      <td>E263</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4012301.0</td>\n",
       "      <td>4.012301e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BK</td>\n",
       "      <td>6429</td>\n",
       "      <td>16</td>\n",
       "      <td>311.0</td>\n",
       "      <td>172.00</td>\n",
       "      <td>2004.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>11228.0</td>\n",
       "      <td>L149</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3017200.0</td>\n",
       "      <td>3.017200e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BK</td>\n",
       "      <td>7790</td>\n",
       "      <td>42</td>\n",
       "      <td>318.0</td>\n",
       "      <td>732.00</td>\n",
       "      <td>1004.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>11210.0</td>\n",
       "      <td>L159</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3073200.0</td>\n",
       "      <td>3.073200e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BK</td>\n",
       "      <td>7197</td>\n",
       "      <td>52</td>\n",
       "      <td>315.0</td>\n",
       "      <td>374.01</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>11223.0</td>\n",
       "      <td>E245</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3037401.0</td>\n",
       "      <td>3.037401e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>QN</td>\n",
       "      <td>15318</td>\n",
       "      <td>35</td>\n",
       "      <td>414.0</td>\n",
       "      <td>1072.01</td>\n",
       "      <td>1032.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>11693.0</td>\n",
       "      <td>E266</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4107201.0</td>\n",
       "      <td>4.107201e+10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 101 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  borough  block  lot  community board  census tract 2010  cb2010  schooldist  \\\n",
       "0      QN    798   31            401.0             123.01  2006.0        30.0   \n",
       "1      BK   6429   16            311.0             172.00  2004.0        20.0   \n",
       "2      BK   7790   42            318.0             732.00  1004.0        22.0   \n",
       "3      BK   7197   52            315.0             374.01  2000.0        21.0   \n",
       "4      QN  15318   35            414.0            1072.01  1032.0        27.0   \n",
       "\n",
       "   council district  postcode firecomp  ...  landmkdate  basempdate  masdate  \\\n",
       "0              22.0   11105.0     E263  ...         NaN         NaN      NaN   \n",
       "1              50.0   11228.0     L149  ...         NaN         NaN      NaN   \n",
       "2              45.0   11210.0     L159  ...         NaN         NaN      NaN   \n",
       "3              47.0   11223.0     E245  ...         NaN         NaN      NaN   \n",
       "4              32.0   11693.0     E266  ...         NaN         NaN      NaN   \n",
       "\n",
       "  polidate edesigdate geom dcpedited notes    bct2020     bctcb2020  \n",
       "0      NaN        NaN  NaN       NaN   NaN  4012301.0  4.012301e+10  \n",
       "1      NaN        NaN  NaN       NaN   NaN  3017200.0  3.017200e+10  \n",
       "2      NaN        NaN  NaN       NaN   NaN  3073200.0  3.073200e+10  \n",
       "3      NaN        NaN  NaN       NaN   NaN  3037401.0  3.037401e+10  \n",
       "4      NaN        NaN  NaN       NaN   NaN  4107201.0  4.107201e+10  \n",
       "\n",
       "[5 rows x 101 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pluto_df['firm07_flag'] = pluto_df['firm07_flag'].fillna(0)\n",
    "pluto_df['pfirm15_flag'] = pluto_df['pfirm15_flag'].fillna(0)\n",
    "\n",
    "pluto_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b049f8b",
   "metadata": {},
   "source": [
    "# Merge datasets into one dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "98e70359",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rows: 644,604\n",
      "columns: 120\n"
     ]
    },
    {
     "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>bin</th>\n",
       "      <th>pluto_bbl</th>\n",
       "      <th>z_floor_grade</th>\n",
       "      <th>the_geom</th>\n",
       "      <th>NAME</th>\n",
       "      <th>BIN</th>\n",
       "      <th>CNSTRCT_YR</th>\n",
       "      <th>LSTMODDATE</th>\n",
       "      <th>LSTSTATYPE</th>\n",
       "      <th>DOITT_ID</th>\n",
       "      <th>...</th>\n",
       "      <th>landmkdate</th>\n",
       "      <th>basempdate</th>\n",
       "      <th>masdate</th>\n",
       "      <th>polidate</th>\n",
       "      <th>edesigdate</th>\n",
       "      <th>geom</th>\n",
       "      <th>dcpedited</th>\n",
       "      <th>notes</th>\n",
       "      <th>bct2020</th>\n",
       "      <th>bctcb2020</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5128004</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>2.904</td>\n",
       "      <td>MULTIPOLYGON (((-74.22271476516362 40.52128189...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5128004</td>\n",
       "      <td>1997.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>651794</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5022601.0</td>\n",
       "      <td>5.022601e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5155392</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>1.337</td>\n",
       "      <td>MULTIPOLYGON (((-74.24172734197434 40.52880752...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5155392</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>1101303</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>5.022602e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5148808</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>2.730</td>\n",
       "      <td>MULTIPOLYGON (((-74.24102137299579 40.52884545...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5148808</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>1101293</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>5.022602e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5088274</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>2.562</td>\n",
       "      <td>MULTIPOLYGON (((-74.24641599963033 40.50699198...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5088274</td>\n",
       "      <td>1980.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>515135</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>5.024401e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5087850</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>4.215</td>\n",
       "      <td>MULTIPOLYGON (((-74.24200532507773 40.51000156...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5087850</td>\n",
       "      <td>1996.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>405357</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>5.024401e+10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 120 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       bin   pluto_bbl  z_floor_grade  \\\n",
       "0  5128004  5075340353          2.904   \n",
       "1  5155392  5075960125          1.337   \n",
       "2  5148808  5075960131          2.730   \n",
       "3  5088274  5079150042          2.562   \n",
       "4  5087850  5078680123          4.215   \n",
       "\n",
       "                                            the_geom NAME      BIN  \\\n",
       "0  MULTIPOLYGON (((-74.22271476516362 40.52128189...  NaN  5128004   \n",
       "1  MULTIPOLYGON (((-74.24172734197434 40.52880752...  NaN  5155392   \n",
       "2  MULTIPOLYGON (((-74.24102137299579 40.52884545...  NaN  5148808   \n",
       "3  MULTIPOLYGON (((-74.24641599963033 40.50699198...  NaN  5088274   \n",
       "4  MULTIPOLYGON (((-74.24200532507773 40.51000156...  NaN  5087850   \n",
       "\n",
       "   CNSTRCT_YR              LSTMODDATE   LSTSTATYPE  DOITT_ID  ...  landmkdate  \\\n",
       "0      1997.0  08/22/2017 12:00:00 AM  Constructed    651794  ...         NaN   \n",
       "1      2007.0  08/22/2017 12:00:00 AM  Constructed   1101303  ...         NaN   \n",
       "2      2007.0  08/22/2017 12:00:00 AM  Constructed   1101293  ...         NaN   \n",
       "3      1980.0  08/22/2017 12:00:00 AM  Constructed    515135  ...         NaN   \n",
       "4      1996.0  08/22/2017 12:00:00 AM  Constructed    405357  ...         NaN   \n",
       "\n",
       "   basempdate  masdate  polidate  edesigdate  geom  dcpedited notes  \\\n",
       "0         NaN      NaN       NaN         NaN   NaN        NaN   NaN   \n",
       "1         NaN      NaN       NaN         NaN   NaN        NaN   NaN   \n",
       "2         NaN      NaN       NaN         NaN   NaN        NaN   NaN   \n",
       "3         NaN      NaN       NaN         NaN   NaN        NaN   NaN   \n",
       "4         NaN      NaN       NaN         NaN   NaN        NaN   NaN   \n",
       "\n",
       "     bct2020     bctcb2020  \n",
       "0  5022601.0  5.022601e+10  \n",
       "1  5022602.0  5.022602e+10  \n",
       "2  5022602.0  5.022602e+10  \n",
       "3  5024401.0  5.024401e+10  \n",
       "4  5024401.0  5.024401e+10  \n",
       "\n",
       "[5 rows x 120 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df = (\n",
    "    bes_df\n",
    "    .loc[:, ['bin', 'pluto_bbl', 'z_floor_grade']]\n",
    "    .merge(\n",
    "        buildings_df,\n",
    "        left_on='bin',\n",
    "        right_on='BIN')\n",
    "    .merge(\n",
    "        pluto_df,\n",
    "        left_on='pluto_bbl',\n",
    "        right_on='bbl')\n",
    ")\n",
    "\n",
    "rows, cols = merged_df.shape\n",
    "print('rows: {:,}\\ncolumns: {}'.format(rows, cols))\n",
    "\n",
    "merged_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "8e23f4d4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rows: 644,604\n",
      "columns: 91\n"
     ]
    },
    {
     "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>bin</th>\n",
       "      <th>pluto_bbl</th>\n",
       "      <th>z_floor_grade</th>\n",
       "      <th>the_geom</th>\n",
       "      <th>BIN</th>\n",
       "      <th>CNSTRCT_YR</th>\n",
       "      <th>LSTMODDATE</th>\n",
       "      <th>LSTSTATYPE</th>\n",
       "      <th>DOITT_ID</th>\n",
       "      <th>HEIGHTROOF</th>\n",
       "      <th>...</th>\n",
       "      <th>sanborn</th>\n",
       "      <th>taxmap</th>\n",
       "      <th>plutomapid</th>\n",
       "      <th>version</th>\n",
       "      <th>sanitdistrict</th>\n",
       "      <th>healthcenterdistrict</th>\n",
       "      <th>firm07_flag</th>\n",
       "      <th>pfirm15_flag</th>\n",
       "      <th>bct2020</th>\n",
       "      <th>bctcb2020</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5128004</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>2.904</td>\n",
       "      <td>MULTIPOLYGON (((-74.22271476516362 40.52128189...</td>\n",
       "      <td>5128004</td>\n",
       "      <td>1997.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>651794</td>\n",
       "      <td>24.683095</td>\n",
       "      <td>...</td>\n",
       "      <td>505 521</td>\n",
       "      <td>53001.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24v1.1</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022601.0</td>\n",
       "      <td>5.022601e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5155392</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>1.337</td>\n",
       "      <td>MULTIPOLYGON (((-74.24172734197434 40.52880752...</td>\n",
       "      <td>5155392</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>1101303</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>505 564</td>\n",
       "      <td>53001.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24v1.1</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>5.022602e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5148808</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>2.730</td>\n",
       "      <td>MULTIPOLYGON (((-74.24102137299579 40.52884545...</td>\n",
       "      <td>5148808</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>1101293</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>505 564</td>\n",
       "      <td>53001.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24v1.1</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>5.022602e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5088274</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>2.562</td>\n",
       "      <td>MULTIPOLYGON (((-74.24641599963033 40.50699198...</td>\n",
       "      <td>5088274</td>\n",
       "      <td>1980.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>515135</td>\n",
       "      <td>24.420000</td>\n",
       "      <td>...</td>\n",
       "      <td>505 531</td>\n",
       "      <td>53201.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24v1.1</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>5.024401e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5087850</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>4.215</td>\n",
       "      <td>MULTIPOLYGON (((-74.24200532507773 40.51000156...</td>\n",
       "      <td>5087850</td>\n",
       "      <td>1996.0</td>\n",
       "      <td>08/22/2017 12:00:00 AM</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>405357</td>\n",
       "      <td>31.410000</td>\n",
       "      <td>...</td>\n",
       "      <td>505 531</td>\n",
       "      <td>53201.0</td>\n",
       "      <td>1</td>\n",
       "      <td>24v1.1</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>5.024401e+10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 91 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       bin   pluto_bbl  z_floor_grade  \\\n",
       "0  5128004  5075340353          2.904   \n",
       "1  5155392  5075960125          1.337   \n",
       "2  5148808  5075960131          2.730   \n",
       "3  5088274  5079150042          2.562   \n",
       "4  5087850  5078680123          4.215   \n",
       "\n",
       "                                            the_geom      BIN  CNSTRCT_YR  \\\n",
       "0  MULTIPOLYGON (((-74.22271476516362 40.52128189...  5128004      1997.0   \n",
       "1  MULTIPOLYGON (((-74.24172734197434 40.52880752...  5155392      2007.0   \n",
       "2  MULTIPOLYGON (((-74.24102137299579 40.52884545...  5148808      2007.0   \n",
       "3  MULTIPOLYGON (((-74.24641599963033 40.50699198...  5088274      1980.0   \n",
       "4  MULTIPOLYGON (((-74.24200532507773 40.51000156...  5087850      1996.0   \n",
       "\n",
       "               LSTMODDATE   LSTSTATYPE  DOITT_ID  HEIGHTROOF  ...  sanborn  \\\n",
       "0  08/22/2017 12:00:00 AM  Constructed    651794   24.683095  ...  505 521   \n",
       "1  08/22/2017 12:00:00 AM  Constructed   1101303   32.000000  ...  505 564   \n",
       "2  08/22/2017 12:00:00 AM  Constructed   1101293   34.000000  ...  505 564   \n",
       "3  08/22/2017 12:00:00 AM  Constructed    515135   24.420000  ...  505 531   \n",
       "4  08/22/2017 12:00:00 AM  Constructed    405357   31.410000  ...  505 531   \n",
       "\n",
       "    taxmap  plutomapid  version  sanitdistrict  healthcenterdistrict  \\\n",
       "0  53001.0           1   24v1.1            3.0                  51.0   \n",
       "1  53001.0           1   24v1.1            3.0                  51.0   \n",
       "2  53001.0           1   24v1.1            3.0                  51.0   \n",
       "3  53201.0           1   24v1.1            3.0                  51.0   \n",
       "4  53201.0           1   24v1.1            3.0                  51.0   \n",
       "\n",
       "  firm07_flag pfirm15_flag    bct2020     bctcb2020  \n",
       "0         0.0          0.0  5022601.0  5.022601e+10  \n",
       "1         0.0          0.0  5022602.0  5.022602e+10  \n",
       "2         0.0          0.0  5022602.0  5.022602e+10  \n",
       "3         0.0          0.0  5024401.0  5.024401e+10  \n",
       "4         0.0          0.0  5024401.0  5.024401e+10  \n",
       "\n",
       "[5 rows x 91 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ser = merged_df.isnull().sum().div(merged_df.shape[0])\n",
    "cols = ser.loc[ser < .50].index.to_list()\n",
    "\n",
    "merged_df = merged_df.loc[:, cols]\n",
    "rows, cols = merged_df.shape\n",
    "print('rows: {:,}\\ncolumns: {}'.format(rows, cols))\n",
    "\n",
    "merged_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "bae84eff",
   "metadata": {},
   "outputs": [
    {
     "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>pluto_bbl</th>\n",
       "      <th>z_floor_grade</th>\n",
       "      <th>CNSTRCT_YR</th>\n",
       "      <th>LSTSTATYPE</th>\n",
       "      <th>HEIGHTROOF</th>\n",
       "      <th>FEAT_CODE</th>\n",
       "      <th>GROUNDELEV</th>\n",
       "      <th>BASE_BBL</th>\n",
       "      <th>MPLUTO_BBL</th>\n",
       "      <th>GEOMSOURCE</th>\n",
       "      <th>...</th>\n",
       "      <th>xcoord</th>\n",
       "      <th>ycoord</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>sanitdistrict</th>\n",
       "      <th>healthcenterdistrict</th>\n",
       "      <th>firm07_flag</th>\n",
       "      <th>pfirm15_flag</th>\n",
       "      <th>bct2020</th>\n",
       "      <th>bctcb2020</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5075340353</td>\n",
       "      <td>2.904</td>\n",
       "      <td>1997.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>24.683095</td>\n",
       "      <td>2100</td>\n",
       "      <td>30.0</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>...</td>\n",
       "      <td>922320.0</td>\n",
       "      <td>129288.0</td>\n",
       "      <td>40.521325</td>\n",
       "      <td>-74.222751</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022601.0</td>\n",
       "      <td>5.022601e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5075960125</td>\n",
       "      <td>1.337</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>2100</td>\n",
       "      <td>29.0</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>Other (Man</td>\n",
       "      <td>...</td>\n",
       "      <td>917024.0</td>\n",
       "      <td>132018.0</td>\n",
       "      <td>40.528779</td>\n",
       "      <td>-74.241827</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>5.022602e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5075960131</td>\n",
       "      <td>2.730</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>2100</td>\n",
       "      <td>34.0</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>Other (Man</td>\n",
       "      <td>...</td>\n",
       "      <td>917221.0</td>\n",
       "      <td>132058.0</td>\n",
       "      <td>40.528891</td>\n",
       "      <td>-74.241118</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>5.022602e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5079150042</td>\n",
       "      <td>2.562</td>\n",
       "      <td>1980.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>24.420000</td>\n",
       "      <td>2100</td>\n",
       "      <td>71.0</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>...</td>\n",
       "      <td>915705.0</td>\n",
       "      <td>124102.0</td>\n",
       "      <td>40.507041</td>\n",
       "      <td>-74.246491</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>5.024401e+10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5078680123</td>\n",
       "      <td>4.215</td>\n",
       "      <td>1996.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>31.410000</td>\n",
       "      <td>2100</td>\n",
       "      <td>75.0</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>...</td>\n",
       "      <td>917021.0</td>\n",
       "      <td>125217.0</td>\n",
       "      <td>40.510112</td>\n",
       "      <td>-74.241770</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>5.024401e+10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 76 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    pluto_bbl  z_floor_grade  CNSTRCT_YR   LSTSTATYPE  HEIGHTROOF  FEAT_CODE  \\\n",
       "0  5075340353          2.904      1997.0  Constructed   24.683095       2100   \n",
       "1  5075960125          1.337      2007.0  Constructed   32.000000       2100   \n",
       "2  5075960131          2.730      2007.0  Constructed   34.000000       2100   \n",
       "3  5079150042          2.562      1980.0  Constructed   24.420000       2100   \n",
       "4  5078680123          4.215      1996.0  Constructed   31.410000       2100   \n",
       "\n",
       "   GROUNDELEV    BASE_BBL  MPLUTO_BBL  GEOMSOURCE  ...    xcoord    ycoord  \\\n",
       "0        30.0  5075340353  5075340353  Photogramm  ...  922320.0  129288.0   \n",
       "1        29.0  5075960125  5075960125  Other (Man  ...  917024.0  132018.0   \n",
       "2        34.0  5075960131  5075960131  Other (Man  ...  917221.0  132058.0   \n",
       "3        71.0  5079150042  5079150042  Photogramm  ...  915705.0  124102.0   \n",
       "4        75.0  5078680123  5078680123  Photogramm  ...  917021.0  125217.0   \n",
       "\n",
       "    latitude  longitude  sanitdistrict  healthcenterdistrict  firm07_flag  \\\n",
       "0  40.521325 -74.222751            3.0                  51.0          0.0   \n",
       "1  40.528779 -74.241827            3.0                  51.0          0.0   \n",
       "2  40.528891 -74.241118            3.0                  51.0          0.0   \n",
       "3  40.507041 -74.246491            3.0                  51.0          0.0   \n",
       "4  40.510112 -74.241770            3.0                  51.0          0.0   \n",
       "\n",
       "   pfirm15_flag    bct2020     bctcb2020  \n",
       "0           0.0  5022601.0  5.022601e+10  \n",
       "1           0.0  5022602.0  5.022602e+10  \n",
       "2           0.0  5022602.0  5.022602e+10  \n",
       "3           0.0  5024401.0  5.024401e+10  \n",
       "4           0.0  5024401.0  5.024401e+10  \n",
       "\n",
       "[5 rows x 76 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols_drop = [\n",
    "    'LSTMODDATE', 'bin', 'BIN', 'the_geom', 'LSTMODDATE', 'DOITT_ID',\n",
    "    'SHAPE_AREA', 'SHAPE_LEN', 'GLOBALID', 'address', 'ownername',\n",
    "    'sanborn', 'taxmap', 'plutomapid', 'version', 'zonemap',\n",
    "]\n",
    "\n",
    "merged_df = merged_df.drop(columns=cols_drop)\n",
    "\n",
    "merged_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "4aa8aac2",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "for col in ['CNSTRCT_YR', 'yearbuilt']:\n",
    "    merged_df.loc[merged_df[col] == 0, col] == np.nan\n",
    "    \n",
    "merged_df.loc[merged_df['yearalter1'] > 0, 'yearalter1'] = 1\n",
    "merged_df.loc[merged_df['yearalter2'] > 0, 'yearalter2'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b660227f",
   "metadata": {},
   "outputs": [
    {
     "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>pluto_bbl</th>\n",
       "      <th>z_floor_grade</th>\n",
       "      <th>CNSTRCT_YR</th>\n",
       "      <th>LSTSTATYPE</th>\n",
       "      <th>HEIGHTROOF</th>\n",
       "      <th>FEAT_CODE</th>\n",
       "      <th>GROUNDELEV</th>\n",
       "      <th>BASE_BBL</th>\n",
       "      <th>MPLUTO_BBL</th>\n",
       "      <th>GEOMSOURCE</th>\n",
       "      <th>...</th>\n",
       "      <th>xcoord</th>\n",
       "      <th>ycoord</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>sanitdistrict</th>\n",
       "      <th>healthcenterdistrict</th>\n",
       "      <th>firm07_flag</th>\n",
       "      <th>pfirm15_flag</th>\n",
       "      <th>bct2020</th>\n",
       "      <th>bctcb2020</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5075340353</td>\n",
       "      <td>2.904</td>\n",
       "      <td>1997.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>24.683095</td>\n",
       "      <td>2100</td>\n",
       "      <td>30.0</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>5075340353</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>...</td>\n",
       "      <td>922320.0</td>\n",
       "      <td>129288.0</td>\n",
       "      <td>40.521325</td>\n",
       "      <td>-74.222751</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022601.0</td>\n",
       "      <td>50226014000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5075960125</td>\n",
       "      <td>1.337</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>2100</td>\n",
       "      <td>29.0</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>5075960125</td>\n",
       "      <td>Other (Man</td>\n",
       "      <td>...</td>\n",
       "      <td>917024.0</td>\n",
       "      <td>132018.0</td>\n",
       "      <td>40.528779</td>\n",
       "      <td>-74.241827</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>50226022005.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5075960131</td>\n",
       "      <td>2.730</td>\n",
       "      <td>2007.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>2100</td>\n",
       "      <td>34.0</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>5075960131</td>\n",
       "      <td>Other (Man</td>\n",
       "      <td>...</td>\n",
       "      <td>917221.0</td>\n",
       "      <td>132058.0</td>\n",
       "      <td>40.528891</td>\n",
       "      <td>-74.241118</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5022602.0</td>\n",
       "      <td>50226022008.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5079150042</td>\n",
       "      <td>2.562</td>\n",
       "      <td>1980.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>24.420000</td>\n",
       "      <td>2100</td>\n",
       "      <td>71.0</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>5079150042</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>...</td>\n",
       "      <td>915705.0</td>\n",
       "      <td>124102.0</td>\n",
       "      <td>40.507041</td>\n",
       "      <td>-74.246491</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>50244011002.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5078680123</td>\n",
       "      <td>4.215</td>\n",
       "      <td>1996.0</td>\n",
       "      <td>Constructed</td>\n",
       "      <td>31.410000</td>\n",
       "      <td>2100</td>\n",
       "      <td>75.0</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>5078680123</td>\n",
       "      <td>Photogramm</td>\n",
       "      <td>...</td>\n",
       "      <td>917021.0</td>\n",
       "      <td>125217.0</td>\n",
       "      <td>40.510112</td>\n",
       "      <td>-74.241770</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5024401.0</td>\n",
       "      <td>50244012005.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 76 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    pluto_bbl  z_floor_grade  CNSTRCT_YR   LSTSTATYPE  HEIGHTROOF FEAT_CODE  \\\n",
       "0  5075340353          2.904      1997.0  Constructed   24.683095      2100   \n",
       "1  5075960125          1.337      2007.0  Constructed   32.000000      2100   \n",
       "2  5075960131          2.730      2007.0  Constructed   34.000000      2100   \n",
       "3  5079150042          2.562      1980.0  Constructed   24.420000      2100   \n",
       "4  5078680123          4.215      1996.0  Constructed   31.410000      2100   \n",
       "\n",
       "   GROUNDELEV    BASE_BBL  MPLUTO_BBL  GEOMSOURCE  ...    xcoord    ycoord  \\\n",
       "0        30.0  5075340353  5075340353  Photogramm  ...  922320.0  129288.0   \n",
       "1        29.0  5075960125  5075960125  Other (Man  ...  917024.0  132018.0   \n",
       "2        34.0  5075960131  5075960131  Other (Man  ...  917221.0  132058.0   \n",
       "3        71.0  5079150042  5079150042  Photogramm  ...  915705.0  124102.0   \n",
       "4        75.0  5078680123  5078680123  Photogramm  ...  917021.0  125217.0   \n",
       "\n",
       "    latitude  longitude sanitdistrict healthcenterdistrict firm07_flag  \\\n",
       "0  40.521325 -74.222751           3.0                 51.0         0.0   \n",
       "1  40.528779 -74.241827           3.0                 51.0         0.0   \n",
       "2  40.528891 -74.241118           3.0                 51.0         0.0   \n",
       "3  40.507041 -74.246491           3.0                 51.0         0.0   \n",
       "4  40.510112 -74.241770           3.0                 51.0         0.0   \n",
       "\n",
       "  pfirm15_flag    bct2020      bctcb2020  \n",
       "0          0.0  5022601.0  50226014000.0  \n",
       "1          0.0  5022602.0  50226022005.0  \n",
       "2          0.0  5022602.0  50226022008.0  \n",
       "3          0.0  5024401.0  50244011002.0  \n",
       "4          0.0  5024401.0  50244012005.0  \n",
       "\n",
       "[5 rows x 76 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dtypes = {\n",
    "    'FEAT_CODE': 'str',\n",
    "    'landuse':'str',\n",
    "    'proxcode':'str',\n",
    "    'lottype':'str',\n",
    "    'bsmtcode':'str',\n",
    "    'firm07_flag':'str',\n",
    "    'pfirm15_flag':'str',\n",
    "    'pluto_bbl':'str',\n",
    "    'BASE_BBL':'str',\n",
    "    'MPLUTO_BBL':'str',\n",
    "    'block':'str',\n",
    "    'lot':'str',\n",
    "    'community board':'str',\n",
    "    'census tract 2010':'str',\n",
    "    'cb2010':'str',\n",
    "    'schooldist':'str',\n",
    "    'council district':'str',\n",
    "    'postcode':'str',\n",
    "    'policeprct':'str',\n",
    "    'sanitboro':'str',\n",
    "    'borocode':'str',\n",
    "    'bbl':'str',\n",
    "    'tract2010':'str',\n",
    "    'sanitdistrict':'str',\n",
    "    'healthcenterdistrict':'str',\n",
    "    'bct2020':'str',\n",
    "    'bctcb2020':'str' ,\n",
    "    'yearalter1':'str',\n",
    "    'yearalter2':'str',\n",
    "}\n",
    "\n",
    "merged_df = merged_df.astype(dtypes)\n",
    "\n",
    "merged_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d7c0c962",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 644604 entries, 0 to 644603\n",
      "Data columns (total 76 columns):\n",
      " #   Column                Non-Null Count   Dtype  \n",
      "---  ------                --------------   -----  \n",
      " 0   pluto_bbl             644604 non-null  object \n",
      " 1   z_floor_grade         644604 non-null  float64\n",
      " 2   CNSTRCT_YR            642492 non-null  float64\n",
      " 3   LSTSTATYPE            644492 non-null  object \n",
      " 4   HEIGHTROOF            643319 non-null  float64\n",
      " 5   FEAT_CODE             644604 non-null  object \n",
      " 6   GROUNDELEV            644385 non-null  float64\n",
      " 7   BASE_BBL              644604 non-null  object \n",
      " 8   MPLUTO_BBL            644604 non-null  object \n",
      " 9   GEOMSOURCE            644525 non-null  object \n",
      " 10  borough               644604 non-null  object \n",
      " 11  block                 644604 non-null  object \n",
      " 12  lot                   644604 non-null  object \n",
      " 13  community board       644604 non-null  object \n",
      " 14  census tract 2010     644604 non-null  object \n",
      " 15  cb2010                644604 non-null  object \n",
      " 16  schooldist            644604 non-null  object \n",
      " 17  council district      644604 non-null  object \n",
      " 18  postcode              644604 non-null  object \n",
      " 19  firecomp              644537 non-null  object \n",
      " 20  policeprct            644604 non-null  object \n",
      " 21  healtharea            644537 non-null  float64\n",
      " 22  sanitboro             644604 non-null  object \n",
      " 23  sanitsub              644222 non-null  object \n",
      " 24  zonedist1             644515 non-null  object \n",
      " 25  splitzone             644515 non-null  object \n",
      " 26  bldgclass             644568 non-null  object \n",
      " 27  landuse               644604 non-null  object \n",
      " 28  easements             644568 non-null  float64\n",
      " 29  lotarea               644568 non-null  float64\n",
      " 30  bldgarea              644567 non-null  float64\n",
      " 31  comarea               640494 non-null  float64\n",
      " 32  resarea               640494 non-null  float64\n",
      " 33  officearea            640494 non-null  float64\n",
      " 34  retailarea            640494 non-null  float64\n",
      " 35  garagearea            640494 non-null  float64\n",
      " 36  strgearea             640494 non-null  float64\n",
      " 37  factryarea            640494 non-null  float64\n",
      " 38  otherarea             640494 non-null  float64\n",
      " 39  areasource            644568 non-null  float64\n",
      " 40  numbldgs              644568 non-null  float64\n",
      " 41  numfloors             641928 non-null  float64\n",
      " 42  unitsres              644566 non-null  float64\n",
      " 43  unitstotal            644567 non-null  float64\n",
      " 44  lotfront              644568 non-null  float64\n",
      " 45  lotdepth              644568 non-null  float64\n",
      " 46  bldgfront             644568 non-null  float64\n",
      " 47  bldgdepth             644568 non-null  float64\n",
      " 48  ext                   620444 non-null  object \n",
      " 49  proxcode              644604 non-null  object \n",
      " 50  irrlotcode            644568 non-null  object \n",
      " 51  lottype               644604 non-null  object \n",
      " 52  bsmtcode              644604 non-null  object \n",
      " 53  assessland            644568 non-null  float64\n",
      " 54  assesstot             644568 non-null  float64\n",
      " 55  exempttot             644568 non-null  float64\n",
      " 56  yearbuilt             644568 non-null  float64\n",
      " 57  yearalter1            644604 non-null  object \n",
      " 58  yearalter2            644604 non-null  object \n",
      " 59  builtfar              644562 non-null  float64\n",
      " 60  residfar              644604 non-null  float64\n",
      " 61  commfar               644604 non-null  float64\n",
      " 62  facilfar              644604 non-null  float64\n",
      " 63  borocode              644604 non-null  object \n",
      " 64  bbl                   644604 non-null  object \n",
      " 65  tract2010             644604 non-null  object \n",
      " 66  xcoord                644540 non-null  float64\n",
      " 67  ycoord                644540 non-null  float64\n",
      " 68  latitude              644540 non-null  float64\n",
      " 69  longitude             644540 non-null  float64\n",
      " 70  sanitdistrict         644604 non-null  object \n",
      " 71  healthcenterdistrict  644604 non-null  object \n",
      " 72  firm07_flag           644604 non-null  object \n",
      " 73  pfirm15_flag          644604 non-null  object \n",
      " 74  bct2020               644604 non-null  object \n",
      " 75  bctcb2020             644604 non-null  object \n",
      "dtypes: float64(37), object(39)\n",
      "memory usage: 378.7+ MB\n"
     ]
    }
   ],
   "source": [
    "merged_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48fbf2f2",
   "metadata": {},
   "source": [
    "# All Land Uses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "d32579e8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 5h 42min 29s, sys: 1min 7s, total: 5h 43min 37s\n",
      "Wall time: 7h 54min 29s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;preprocess&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;cat&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  [&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;,\n",
       "                                                   &#x27;FEAT_CODE&#x27;, &#x27;BASE_BBL&#x27;,\n",
       "                                                   &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;,\n",
       "                                                   &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;,\n",
       "                                                   &#x27;community board&#x27;,\n",
       "                                                   &#x27;census tract 2010&#x27;,\n",
       "                                                   &#x27;cb2010&#x27;, &#x27;schooldist&#x27;,\n",
       "                                                   &#x27;council district&#x27;,\n",
       "                                                   &#x27;postcode&#x27;, &#x27;firecomp&#x27;,\n",
       "                                                   &#x27;policeprct&#x27;, &#x27;sa...\n",
       "                                                   &#x27;retailarea&#x27;, &#x27;garagearea&#x27;,\n",
       "                                                   &#x27;strgearea&#x27;, &#x27;factryarea&#x27;,\n",
       "                                                   &#x27;otherarea&#x27;, &#x27;areasource&#x27;,\n",
       "                                                   &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;,\n",
       "                                                   &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;,\n",
       "                                                   &#x27;lotfront&#x27;, &#x27;lotdepth&#x27;,\n",
       "                                                   &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;,\n",
       "                                                   &#x27;assessland&#x27;, &#x27;assesstot&#x27;,\n",
       "                                                   &#x27;exempttot&#x27;, &#x27;yearbuilt&#x27;,\n",
       "                                                   &#x27;builtfar&#x27;, &#x27;residfar&#x27;, ...])],\n",
       "                                   verbose_feature_names_out=False)),\n",
       "                (&#x27;classifier&#x27;,\n",
       "                 RandomForestRegressor(max_depth=10, n_jobs=-1,\n",
       "                                       random_state=42))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;preprocess&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;cat&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  [&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;,\n",
       "                                                   &#x27;FEAT_CODE&#x27;, &#x27;BASE_BBL&#x27;,\n",
       "                                                   &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;,\n",
       "                                                   &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;,\n",
       "                                                   &#x27;community board&#x27;,\n",
       "                                                   &#x27;census tract 2010&#x27;,\n",
       "                                                   &#x27;cb2010&#x27;, &#x27;schooldist&#x27;,\n",
       "                                                   &#x27;council district&#x27;,\n",
       "                                                   &#x27;postcode&#x27;, &#x27;firecomp&#x27;,\n",
       "                                                   &#x27;policeprct&#x27;, &#x27;sa...\n",
       "                                                   &#x27;retailarea&#x27;, &#x27;garagearea&#x27;,\n",
       "                                                   &#x27;strgearea&#x27;, &#x27;factryarea&#x27;,\n",
       "                                                   &#x27;otherarea&#x27;, &#x27;areasource&#x27;,\n",
       "                                                   &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;,\n",
       "                                                   &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;,\n",
       "                                                   &#x27;lotfront&#x27;, &#x27;lotdepth&#x27;,\n",
       "                                                   &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;,\n",
       "                                                   &#x27;assessland&#x27;, &#x27;assesstot&#x27;,\n",
       "                                                   &#x27;exempttot&#x27;, &#x27;yearbuilt&#x27;,\n",
       "                                                   &#x27;builtfar&#x27;, &#x27;residfar&#x27;, ...])],\n",
       "                                   verbose_feature_names_out=False)),\n",
       "                (&#x27;classifier&#x27;,\n",
       "                 RandomForestRegressor(max_depth=10, n_jobs=-1,\n",
       "                                       random_state=42))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">preprocess: ColumnTransformer</label><div class=\"sk-toggleable__content\"><pre>ColumnTransformer(transformers=[(&#x27;cat&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                 [&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;, &#x27;FEAT_CODE&#x27;,\n",
       "                                  &#x27;BASE_BBL&#x27;, &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;,\n",
       "                                  &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;, &#x27;community board&#x27;,\n",
       "                                  &#x27;census tract 2010&#x27;, &#x27;cb2010&#x27;, &#x27;schooldist&#x27;,\n",
       "                                  &#x27;council district&#x27;, &#x27;postcode&#x27;, &#x27;firecomp&#x27;,\n",
       "                                  &#x27;policeprct&#x27;, &#x27;sanitboro&#x27;, &#x27;sanitsub&#x27;,\n",
       "                                  &#x27;zonedist1...\n",
       "                                  &#x27;healtharea&#x27;, &#x27;easements&#x27;, &#x27;lotarea&#x27;,\n",
       "                                  &#x27;bldgarea&#x27;, &#x27;comarea&#x27;, &#x27;resarea&#x27;,\n",
       "                                  &#x27;officearea&#x27;, &#x27;retailarea&#x27;, &#x27;garagearea&#x27;,\n",
       "                                  &#x27;strgearea&#x27;, &#x27;factryarea&#x27;, &#x27;otherarea&#x27;,\n",
       "                                  &#x27;areasource&#x27;, &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;,\n",
       "                                  &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;, &#x27;lotfront&#x27;,\n",
       "                                  &#x27;lotdepth&#x27;, &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;,\n",
       "                                  &#x27;assessland&#x27;, &#x27;assesstot&#x27;, &#x27;exempttot&#x27;,\n",
       "                                  &#x27;yearbuilt&#x27;, &#x27;builtfar&#x27;, &#x27;residfar&#x27;, ...])],\n",
       "                  verbose_feature_names_out=False)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">cat</label><div class=\"sk-toggleable__content\"><pre>[&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;, &#x27;FEAT_CODE&#x27;, &#x27;BASE_BBL&#x27;, &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;, &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;, &#x27;community board&#x27;, &#x27;census tract 2010&#x27;, &#x27;cb2010&#x27;, &#x27;schooldist&#x27;, &#x27;council district&#x27;, &#x27;postcode&#x27;, &#x27;firecomp&#x27;, &#x27;policeprct&#x27;, &#x27;sanitboro&#x27;, &#x27;sanitsub&#x27;, &#x27;zonedist1&#x27;, &#x27;splitzone&#x27;, &#x27;bldgclass&#x27;, &#x27;landuse&#x27;, &#x27;ext&#x27;, &#x27;proxcode&#x27;, &#x27;irrlotcode&#x27;, &#x27;lottype&#x27;, &#x27;bsmtcode&#x27;, &#x27;yearalter1&#x27;, &#x27;yearalter2&#x27;, &#x27;borocode&#x27;, &#x27;bbl&#x27;, &#x27;tract2010&#x27;, &#x27;sanitdistrict&#x27;, &#x27;healthcenterdistrict&#x27;, &#x27;firm07_flag&#x27;, &#x27;pfirm15_flag&#x27;, &#x27;bct2020&#x27;, &#x27;bctcb2020&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">OneHotEncoder</label><div class=\"sk-toggleable__content\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div></div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-5\" type=\"checkbox\" ><label for=\"sk-estimator-id-5\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">num</label><div class=\"sk-toggleable__content\"><pre>[&#x27;CNSTRCT_YR&#x27;, &#x27;HEIGHTROOF&#x27;, &#x27;GROUNDELEV&#x27;, &#x27;healtharea&#x27;, &#x27;easements&#x27;, &#x27;lotarea&#x27;, &#x27;bldgarea&#x27;, &#x27;comarea&#x27;, &#x27;resarea&#x27;, &#x27;officearea&#x27;, &#x27;retailarea&#x27;, &#x27;garagearea&#x27;, &#x27;strgearea&#x27;, &#x27;factryarea&#x27;, &#x27;otherarea&#x27;, &#x27;areasource&#x27;, &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;, &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;, &#x27;lotfront&#x27;, &#x27;lotdepth&#x27;, &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;, &#x27;assessland&#x27;, &#x27;assesstot&#x27;, &#x27;exempttot&#x27;, &#x27;yearbuilt&#x27;, &#x27;builtfar&#x27;, &#x27;residfar&#x27;, &#x27;commfar&#x27;, &#x27;facilfar&#x27;, &#x27;xcoord&#x27;, &#x27;ycoord&#x27;, &#x27;latitude&#x27;, &#x27;longitude&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-6\" type=\"checkbox\" ><label for=\"sk-estimator-id-6\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">SimpleImputer</label><div class=\"sk-toggleable__content\"><pre>SimpleImputer()</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-7\" type=\"checkbox\" ><label for=\"sk-estimator-id-7\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">StandardScaler</label><div class=\"sk-toggleable__content\"><pre>StandardScaler()</pre></div></div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-8\" type=\"checkbox\" ><label for=\"sk-estimator-id-8\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomForestRegressor</label><div class=\"sk-toggleable__content\"><pre>RandomForestRegressor(max_depth=10, n_jobs=-1, random_state=42)</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('preprocess',\n",
       "                 ColumnTransformer(transformers=[('cat',\n",
       "                                                  Pipeline(steps=[('encoder',\n",
       "                                                                   OneHotEncoder(handle_unknown='ignore'))]),\n",
       "                                                  ['pluto_bbl', 'LSTSTATYPE',\n",
       "                                                   'FEAT_CODE', 'BASE_BBL',\n",
       "                                                   'MPLUTO_BBL', 'GEOMSOURCE',\n",
       "                                                   'borough', 'block', 'lot',\n",
       "                                                   'community board',\n",
       "                                                   'census tract 2010',\n",
       "                                                   'cb2010', 'schooldist',\n",
       "                                                   'council district',\n",
       "                                                   'postcode', 'firecomp',\n",
       "                                                   'policeprct', 'sa...\n",
       "                                                   'retailarea', 'garagearea',\n",
       "                                                   'strgearea', 'factryarea',\n",
       "                                                   'otherarea', 'areasource',\n",
       "                                                   'numbldgs', 'numfloors',\n",
       "                                                   'unitsres', 'unitstotal',\n",
       "                                                   'lotfront', 'lotdepth',\n",
       "                                                   'bldgfront', 'bldgdepth',\n",
       "                                                   'assessland', 'assesstot',\n",
       "                                                   'exempttot', 'yearbuilt',\n",
       "                                                   'builtfar', 'residfar', ...])],\n",
       "                                   verbose_feature_names_out=False)),\n",
       "                ('classifier',\n",
       "                 RandomForestRegressor(max_depth=10, n_jobs=-1,\n",
       "                                       random_state=42))])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "categorical_columns = (\n",
    "    merged_df\n",
    "    .select_dtypes(exclude='number')\n",
    "    .columns\n",
    "    .to_list()\n",
    ")\n",
    "\n",
    "numerical_columns = (\n",
    "    merged_df\n",
    "    .drop(columns=['z_floor_grade'])\n",
    "    .select_dtypes(include='number')\n",
    "    .columns\n",
    "    .to_list()\n",
    ")\n",
    "\n",
    "X = merged_df[numerical_columns + categorical_columns]\n",
    "y = merged_df['z_floor_grade']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)\n",
    "\n",
    "categorical_transformer = Pipeline(\n",
    "    steps=[(\"encoder\", OneHotEncoder(handle_unknown=\"ignore\"))]\n",
    ")\n",
    "numeric_transformer = Pipeline(\n",
    "    steps=[(\"imputer\", SimpleImputer(strategy=\"mean\")), (\"scaler\", StandardScaler())]\n",
    ")\n",
    "\n",
    "preprocessing = ColumnTransformer([\n",
    "        (\"cat\", categorical_transformer, categorical_columns),\n",
    "        (\"num\", numeric_transformer, numerical_columns)],\n",
    "    verbose_feature_names_out=False,\n",
    ")\n",
    "\n",
    "rf = Pipeline([\n",
    "    (\"preprocess\", preprocessing),\n",
    "    (\"classifier\", RandomForestRegressor(random_state=42, max_depth=10, n_jobs=-1))]\n",
    ")\n",
    "\n",
    "rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "1e2ec5f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RF train accuracy r-squared: 0.396\n",
      "RF test accuracy r-squared: 0.367\n",
      "RMSE: 2.298\n"
     ]
    }
   ],
   "source": [
    "print(f\"RF train accuracy r-squared: {rf.score(X_train, y_train):.3f}\")\n",
    "print(f\"RF test accuracy r-squared: {rf.score(X_test, y_test):.3f}\")\n",
    "print(f\"RMSE: {mean_squared_error(y_test, rf.predict(X_test), squared=False):.3f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "4ecd88c2",
   "metadata": {},
   "outputs": [
    {
     "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>y_test</th>\n",
       "      <th>y_pred</th>\n",
       "      <th>residuals</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.112</td>\n",
       "      <td>3.355149</td>\n",
       "      <td>0.756851</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8.009</td>\n",
       "      <td>7.732464</td>\n",
       "      <td>0.276536</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8.275</td>\n",
       "      <td>7.396516</td>\n",
       "      <td>0.878484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.985</td>\n",
       "      <td>3.753995</td>\n",
       "      <td>-0.768995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.188</td>\n",
       "      <td>3.246289</td>\n",
       "      <td>0.941711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161146</th>\n",
       "      <td>5.422</td>\n",
       "      <td>5.569909</td>\n",
       "      <td>-0.147909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161147</th>\n",
       "      <td>2.770</td>\n",
       "      <td>4.239332</td>\n",
       "      <td>-1.469332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161148</th>\n",
       "      <td>9.436</td>\n",
       "      <td>4.390654</td>\n",
       "      <td>5.045346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161149</th>\n",
       "      <td>2.839</td>\n",
       "      <td>3.241539</td>\n",
       "      <td>-0.402539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161150</th>\n",
       "      <td>7.687</td>\n",
       "      <td>4.275938</td>\n",
       "      <td>3.411062</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>161151 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        y_test    y_pred  residuals\n",
       "0        4.112  3.355149   0.756851\n",
       "1        8.009  7.732464   0.276536\n",
       "2        8.275  7.396516   0.878484\n",
       "3        2.985  3.753995  -0.768995\n",
       "4        4.188  3.246289   0.941711\n",
       "...        ...       ...        ...\n",
       "161146   5.422  5.569909  -0.147909\n",
       "161147   2.770  4.239332  -1.469332\n",
       "161148   9.436  4.390654   5.045346\n",
       "161149   2.839  3.241539  -0.402539\n",
       "161150   7.687  4.275938   3.411062\n",
       "\n",
       "[161151 rows x 3 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = pd.Series(rf.predict(X_test), name='y_pred')\n",
    "evaluate_df = pd.concat(\n",
    "    [y_test.reset_index(drop=True).rename('y_test'), y_pred],\n",
    "    axis=1\n",
    ")\n",
    "\n",
    "evaluate_df['residuals'] = evaluate_df['y_test'] - evaluate_df['y_pred']\n",
    "\n",
    "evaluate_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "ff8579f0",
   "metadata": {},
   "outputs": [
    {
     "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>y_test</th>\n",
       "      <th>y_pred</th>\n",
       "      <th>residuals</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>161151.000000</td>\n",
       "      <td>161151.000000</td>\n",
       "      <td>161151.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>3.831730</td>\n",
       "      <td>3.836050</td>\n",
       "      <td>-0.004320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2.888553</td>\n",
       "      <td>1.673607</td>\n",
       "      <td>2.297902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-8.832000</td>\n",
       "      <td>0.173868</td>\n",
       "      <td>-12.633158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.671000</td>\n",
       "      <td>3.213240</td>\n",
       "      <td>-1.053635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>3.363000</td>\n",
       "      <td>3.676253</td>\n",
       "      <td>-0.198240</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>4.988000</td>\n",
       "      <td>4.319250</td>\n",
       "      <td>0.793186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>82.511000</td>\n",
       "      <td>15.582396</td>\n",
       "      <td>67.236788</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              y_test         y_pred      residuals\n",
       "count  161151.000000  161151.000000  161151.000000\n",
       "mean        3.831730       3.836050      -0.004320\n",
       "std         2.888553       1.673607       2.297902\n",
       "min        -8.832000       0.173868     -12.633158\n",
       "25%         1.671000       3.213240      -1.053635\n",
       "50%         3.363000       3.676253      -0.198240\n",
       "75%         4.988000       4.319250       0.793186\n",
       "max        82.511000      15.582396      67.236788"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "evaluate_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "d82253a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(5, 5))\n",
    "\n",
    "sns.scatterplot(\n",
    "    data=evaluate_df,\n",
    "    y='y_test',\n",
    "    x='y_pred',\n",
    "    ax=ax,\n",
    "    alpha=.3\n",
    ")\n",
    "\n",
    "right = evaluate_df.iloc[:,:-1].max().max() + 3\n",
    "left = evaluate_df.iloc[:,:-1].min().min() - 3\n",
    "ax.set_xlim(left, right)\n",
    "ax.set_ylim(left, right)\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d5a74ad3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(5, 5))\n",
    "\n",
    "sns.scatterplot(\n",
    "    data=evaluate_df,\n",
    "    y='residuals',\n",
    "    x='y_pred',\n",
    "    ax=ax,\n",
    "    alpha=.3\n",
    ")\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f5b15b9c",
   "metadata": {},
   "source": [
    "## All Land Uses: Feature importance based on Mean Decrease in Impurity (MDI)\n",
    "Source: https://scikit-learn.org/stable/auto_examples/inspection/plot_permutation_importance.html#sphx-glr-auto-examples-inspection-plot-permutation-importance-py\n",
    "\n",
    "Further reading: https://scikit-learn.org/stable/auto_examples/ensemble/plot_forest_importances.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "4b987074",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "bldgclass_A5    0.004721\n",
       "bsmtcode_2.0    0.005350\n",
       "commfar         0.005686\n",
       "ycoord          0.007042\n",
       "ext_G           0.007089\n",
       "lotfront        0.009092\n",
       "latitude        0.009649\n",
       "builtfar        0.010524\n",
       "resarea         0.011208\n",
       "bsmtcode_5.0    0.011616\n",
       "CNSTRCT_YR      0.015426\n",
       "yearbuilt       0.016950\n",
       "lotarea         0.017678\n",
       "numbldgs        0.018137\n",
       "retailarea      0.019019\n",
       "healtharea      0.022230\n",
       "numfloors       0.038822\n",
       "HEIGHTROOF      0.063234\n",
       "comarea         0.180547\n",
       "bsmtcode_1.0    0.367274\n",
       "dtype: float64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_names = rf[:-1].get_feature_names_out()\n",
    "\n",
    "mdi_importances = (\n",
    "    pd\n",
    "    .Series(rf[-1].feature_importances_, index=feature_names)\n",
    "    .sort_values(ascending=True)\n",
    "    .tail(20)\n",
    ")\n",
    "\n",
    "mdi_importances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "ab82e5e2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8, 6))\n",
    "\n",
    "mdi_importances.plot.barh(ax=ax)\n",
    "\n",
    "plt.title(\"Random Forest Feature Importances (MDI)\")\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab648158",
   "metadata": {},
   "source": [
    "# Residential Land Use Only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "435d001a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rows: 557,925\n",
      "columns: 76\n"
     ]
    }
   ],
   "source": [
    "residential_df = (\n",
    "    merged_df\n",
    "    .loc[merged_df['landuse'].isin(['1.0', '2.0', '3.0'])]\n",
    "    .reset_index(drop=True)\n",
    ")\n",
    "\n",
    "rows, cols = residential_df.shape\n",
    "print('rows: {:,}\\ncolumns: {}'.format(rows, cols))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "41bdfb07",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3h 5min 52s, sys: 20.7 s, total: 3h 6min 13s\n",
      "Wall time: 57min 9s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-2 {color: black;background-color: white;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;preprocess&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;cat&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  [&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;,\n",
       "                                                   &#x27;FEAT_CODE&#x27;, &#x27;BASE_BBL&#x27;,\n",
       "                                                   &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;,\n",
       "                                                   &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;,\n",
       "                                                   &#x27;community board&#x27;,\n",
       "                                                   &#x27;census tract 2010&#x27;,\n",
       "                                                   &#x27;cb2010&#x27;, &#x27;schooldist&#x27;,\n",
       "                                                   &#x27;council district&#x27;,\n",
       "                                                   &#x27;postcode&#x27;, &#x27;firecomp&#x27;,\n",
       "                                                   &#x27;policeprct&#x27;, &#x27;sa...\n",
       "                                                   &#x27;retailarea&#x27;, &#x27;garagearea&#x27;,\n",
       "                                                   &#x27;strgearea&#x27;, &#x27;factryarea&#x27;,\n",
       "                                                   &#x27;otherarea&#x27;, &#x27;areasource&#x27;,\n",
       "                                                   &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;,\n",
       "                                                   &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;,\n",
       "                                                   &#x27;lotfront&#x27;, &#x27;lotdepth&#x27;,\n",
       "                                                   &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;,\n",
       "                                                   &#x27;assessland&#x27;, &#x27;assesstot&#x27;,\n",
       "                                                   &#x27;exempttot&#x27;, &#x27;yearbuilt&#x27;,\n",
       "                                                   &#x27;builtfar&#x27;, &#x27;residfar&#x27;, ...])],\n",
       "                                   verbose_feature_names_out=False)),\n",
       "                (&#x27;classifier&#x27;,\n",
       "                 RandomForestRegressor(max_depth=10, n_jobs=-1,\n",
       "                                       random_state=42))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-9\" type=\"checkbox\" ><label for=\"sk-estimator-id-9\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;preprocess&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;cat&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                                   OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                                  [&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;,\n",
       "                                                   &#x27;FEAT_CODE&#x27;, &#x27;BASE_BBL&#x27;,\n",
       "                                                   &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;,\n",
       "                                                   &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;,\n",
       "                                                   &#x27;community board&#x27;,\n",
       "                                                   &#x27;census tract 2010&#x27;,\n",
       "                                                   &#x27;cb2010&#x27;, &#x27;schooldist&#x27;,\n",
       "                                                   &#x27;council district&#x27;,\n",
       "                                                   &#x27;postcode&#x27;, &#x27;firecomp&#x27;,\n",
       "                                                   &#x27;policeprct&#x27;, &#x27;sa...\n",
       "                                                   &#x27;retailarea&#x27;, &#x27;garagearea&#x27;,\n",
       "                                                   &#x27;strgearea&#x27;, &#x27;factryarea&#x27;,\n",
       "                                                   &#x27;otherarea&#x27;, &#x27;areasource&#x27;,\n",
       "                                                   &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;,\n",
       "                                                   &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;,\n",
       "                                                   &#x27;lotfront&#x27;, &#x27;lotdepth&#x27;,\n",
       "                                                   &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;,\n",
       "                                                   &#x27;assessland&#x27;, &#x27;assesstot&#x27;,\n",
       "                                                   &#x27;exempttot&#x27;, &#x27;yearbuilt&#x27;,\n",
       "                                                   &#x27;builtfar&#x27;, &#x27;residfar&#x27;, ...])],\n",
       "                                   verbose_feature_names_out=False)),\n",
       "                (&#x27;classifier&#x27;,\n",
       "                 RandomForestRegressor(max_depth=10, n_jobs=-1,\n",
       "                                       random_state=42))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-10\" type=\"checkbox\" ><label for=\"sk-estimator-id-10\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">preprocess: ColumnTransformer</label><div class=\"sk-toggleable__content\"><pre>ColumnTransformer(transformers=[(&#x27;cat&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;encoder&#x27;,\n",
       "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
       "                                 [&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;, &#x27;FEAT_CODE&#x27;,\n",
       "                                  &#x27;BASE_BBL&#x27;, &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;,\n",
       "                                  &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;, &#x27;community board&#x27;,\n",
       "                                  &#x27;census tract 2010&#x27;, &#x27;cb2010&#x27;, &#x27;schooldist&#x27;,\n",
       "                                  &#x27;council district&#x27;, &#x27;postcode&#x27;, &#x27;firecomp&#x27;,\n",
       "                                  &#x27;policeprct&#x27;, &#x27;sanitboro&#x27;, &#x27;sanitsub&#x27;,\n",
       "                                  &#x27;zonedist1...\n",
       "                                  &#x27;healtharea&#x27;, &#x27;easements&#x27;, &#x27;lotarea&#x27;,\n",
       "                                  &#x27;bldgarea&#x27;, &#x27;comarea&#x27;, &#x27;resarea&#x27;,\n",
       "                                  &#x27;officearea&#x27;, &#x27;retailarea&#x27;, &#x27;garagearea&#x27;,\n",
       "                                  &#x27;strgearea&#x27;, &#x27;factryarea&#x27;, &#x27;otherarea&#x27;,\n",
       "                                  &#x27;areasource&#x27;, &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;,\n",
       "                                  &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;, &#x27;lotfront&#x27;,\n",
       "                                  &#x27;lotdepth&#x27;, &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;,\n",
       "                                  &#x27;assessland&#x27;, &#x27;assesstot&#x27;, &#x27;exempttot&#x27;,\n",
       "                                  &#x27;yearbuilt&#x27;, &#x27;builtfar&#x27;, &#x27;residfar&#x27;, ...])],\n",
       "                  verbose_feature_names_out=False)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-11\" type=\"checkbox\" ><label for=\"sk-estimator-id-11\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">cat</label><div class=\"sk-toggleable__content\"><pre>[&#x27;pluto_bbl&#x27;, &#x27;LSTSTATYPE&#x27;, &#x27;FEAT_CODE&#x27;, &#x27;BASE_BBL&#x27;, &#x27;MPLUTO_BBL&#x27;, &#x27;GEOMSOURCE&#x27;, &#x27;borough&#x27;, &#x27;block&#x27;, &#x27;lot&#x27;, &#x27;community board&#x27;, &#x27;census tract 2010&#x27;, &#x27;cb2010&#x27;, &#x27;schooldist&#x27;, &#x27;council district&#x27;, &#x27;postcode&#x27;, &#x27;firecomp&#x27;, &#x27;policeprct&#x27;, &#x27;sanitboro&#x27;, &#x27;sanitsub&#x27;, &#x27;zonedist1&#x27;, &#x27;splitzone&#x27;, &#x27;bldgclass&#x27;, &#x27;landuse&#x27;, &#x27;ext&#x27;, &#x27;proxcode&#x27;, &#x27;irrlotcode&#x27;, &#x27;lottype&#x27;, &#x27;bsmtcode&#x27;, &#x27;yearalter1&#x27;, &#x27;yearalter2&#x27;, &#x27;borocode&#x27;, &#x27;bbl&#x27;, &#x27;tract2010&#x27;, &#x27;sanitdistrict&#x27;, &#x27;healthcenterdistrict&#x27;, &#x27;firm07_flag&#x27;, &#x27;pfirm15_flag&#x27;, &#x27;bct2020&#x27;, &#x27;bctcb2020&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-12\" type=\"checkbox\" ><label for=\"sk-estimator-id-12\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">OneHotEncoder</label><div class=\"sk-toggleable__content\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div></div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-13\" type=\"checkbox\" ><label for=\"sk-estimator-id-13\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">num</label><div class=\"sk-toggleable__content\"><pre>[&#x27;CNSTRCT_YR&#x27;, &#x27;HEIGHTROOF&#x27;, &#x27;GROUNDELEV&#x27;, &#x27;healtharea&#x27;, &#x27;easements&#x27;, &#x27;lotarea&#x27;, &#x27;bldgarea&#x27;, &#x27;comarea&#x27;, &#x27;resarea&#x27;, &#x27;officearea&#x27;, &#x27;retailarea&#x27;, &#x27;garagearea&#x27;, &#x27;strgearea&#x27;, &#x27;factryarea&#x27;, &#x27;otherarea&#x27;, &#x27;areasource&#x27;, &#x27;numbldgs&#x27;, &#x27;numfloors&#x27;, &#x27;unitsres&#x27;, &#x27;unitstotal&#x27;, &#x27;lotfront&#x27;, &#x27;lotdepth&#x27;, &#x27;bldgfront&#x27;, &#x27;bldgdepth&#x27;, &#x27;assessland&#x27;, &#x27;assesstot&#x27;, &#x27;exempttot&#x27;, &#x27;yearbuilt&#x27;, &#x27;builtfar&#x27;, &#x27;residfar&#x27;, &#x27;commfar&#x27;, &#x27;facilfar&#x27;, &#x27;xcoord&#x27;, &#x27;ycoord&#x27;, &#x27;latitude&#x27;, &#x27;longitude&#x27;]</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-14\" type=\"checkbox\" ><label for=\"sk-estimator-id-14\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">SimpleImputer</label><div class=\"sk-toggleable__content\"><pre>SimpleImputer()</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-15\" type=\"checkbox\" ><label for=\"sk-estimator-id-15\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">StandardScaler</label><div class=\"sk-toggleable__content\"><pre>StandardScaler()</pre></div></div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-16\" type=\"checkbox\" ><label for=\"sk-estimator-id-16\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomForestRegressor</label><div class=\"sk-toggleable__content\"><pre>RandomForestRegressor(max_depth=10, n_jobs=-1, random_state=42)</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('preprocess',\n",
       "                 ColumnTransformer(transformers=[('cat',\n",
       "                                                  Pipeline(steps=[('encoder',\n",
       "                                                                   OneHotEncoder(handle_unknown='ignore'))]),\n",
       "                                                  ['pluto_bbl', 'LSTSTATYPE',\n",
       "                                                   'FEAT_CODE', 'BASE_BBL',\n",
       "                                                   'MPLUTO_BBL', 'GEOMSOURCE',\n",
       "                                                   'borough', 'block', 'lot',\n",
       "                                                   'community board',\n",
       "                                                   'census tract 2010',\n",
       "                                                   'cb2010', 'schooldist',\n",
       "                                                   'council district',\n",
       "                                                   'postcode', 'firecomp',\n",
       "                                                   'policeprct', 'sa...\n",
       "                                                   'retailarea', 'garagearea',\n",
       "                                                   'strgearea', 'factryarea',\n",
       "                                                   'otherarea', 'areasource',\n",
       "                                                   'numbldgs', 'numfloors',\n",
       "                                                   'unitsres', 'unitstotal',\n",
       "                                                   'lotfront', 'lotdepth',\n",
       "                                                   'bldgfront', 'bldgdepth',\n",
       "                                                   'assessland', 'assesstot',\n",
       "                                                   'exempttot', 'yearbuilt',\n",
       "                                                   'builtfar', 'residfar', ...])],\n",
       "                                   verbose_feature_names_out=False)),\n",
       "                ('classifier',\n",
       "                 RandomForestRegressor(max_depth=10, n_jobs=-1,\n",
       "                                       random_state=42))])"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "categorical_columns = (\n",
    "    residential_df\n",
    "    .select_dtypes(exclude='number')\n",
    "    .columns\n",
    "    .to_list()\n",
    ")\n",
    "\n",
    "numerical_columns = (\n",
    "    residential_df\n",
    "    .drop(columns=['z_floor_grade'])\n",
    "    .select_dtypes(include='number')\n",
    "    .columns\n",
    "    .to_list()\n",
    ")\n",
    "\n",
    "X = residential_df[numerical_columns + categorical_columns]\n",
    "y = residential_df['z_floor_grade']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)\n",
    "\n",
    "categorical_transformer = Pipeline(\n",
    "    steps=[(\"encoder\", OneHotEncoder(handle_unknown=\"ignore\"))]\n",
    ")\n",
    "numeric_transformer = Pipeline(\n",
    "    steps=[(\"imputer\", SimpleImputer(strategy=\"mean\")), (\"scaler\", StandardScaler())]\n",
    ")\n",
    "\n",
    "preprocessing = ColumnTransformer([\n",
    "        (\"cat\", categorical_transformer, categorical_columns),\n",
    "        (\"num\", numeric_transformer, numerical_columns)],\n",
    "    verbose_feature_names_out=False,\n",
    ")\n",
    "\n",
    "rf = Pipeline([\n",
    "    (\"preprocess\", preprocessing),\n",
    "    (\"classifier\", RandomForestRegressor(random_state=42, max_depth=10, n_jobs=-1))]\n",
    ")\n",
    "\n",
    "rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "5386b55d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RF train accuracy r-squared: 0.349\n",
      "RF test accuracy r-squared: 0.321\n",
      "RMSE: 2.277\n"
     ]
    }
   ],
   "source": [
    "print(f\"RF train accuracy r-squared: {rf.score(X_train, y_train):.3f}\")\n",
    "print(f\"RF test accuracy r-squared: {rf.score(X_test, y_test):.3f}\")\n",
    "print(f\"RMSE: {mean_squared_error(y_test, rf.predict(X_test), squared=False):.3f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "a2fd099d",
   "metadata": {},
   "outputs": [
    {
     "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>y_test</th>\n",
       "      <th>y_pred</th>\n",
       "      <th>residuals</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.224</td>\n",
       "      <td>3.029089</td>\n",
       "      <td>-2.805089</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3.519</td>\n",
       "      <td>3.508187</td>\n",
       "      <td>0.010813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9.201</td>\n",
       "      <td>1.418478</td>\n",
       "      <td>7.782522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.086</td>\n",
       "      <td>3.277603</td>\n",
       "      <td>-0.191603</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.817</td>\n",
       "      <td>3.724338</td>\n",
       "      <td>0.092662</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139477</th>\n",
       "      <td>3.284</td>\n",
       "      <td>4.202190</td>\n",
       "      <td>-0.918190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139478</th>\n",
       "      <td>4.057</td>\n",
       "      <td>5.275799</td>\n",
       "      <td>-1.218799</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139479</th>\n",
       "      <td>3.290</td>\n",
       "      <td>3.273910</td>\n",
       "      <td>0.016090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139480</th>\n",
       "      <td>2.914</td>\n",
       "      <td>2.676162</td>\n",
       "      <td>0.237838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139481</th>\n",
       "      <td>3.016</td>\n",
       "      <td>3.412853</td>\n",
       "      <td>-0.396853</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>139482 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        y_test    y_pred  residuals\n",
       "0        0.224  3.029089  -2.805089\n",
       "1        3.519  3.508187   0.010813\n",
       "2        9.201  1.418478   7.782522\n",
       "3        3.086  3.277603  -0.191603\n",
       "4        3.817  3.724338   0.092662\n",
       "...        ...       ...        ...\n",
       "139477   3.284  4.202190  -0.918190\n",
       "139478   4.057  5.275799  -1.218799\n",
       "139479   3.290  3.273910   0.016090\n",
       "139480   2.914  2.676162   0.237838\n",
       "139481   3.016  3.412853  -0.396853\n",
       "\n",
       "[139482 rows x 3 columns]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = pd.Series(rf.predict(X_test), name='y_pred')\n",
    "evaluate_df = pd.concat(\n",
    "    [y_test.reset_index(drop=True).rename('y_test'), y_pred],\n",
    "    axis=1\n",
    ")\n",
    "\n",
    "evaluate_df['residuals'] = evaluate_df['y_test'] - evaluate_df['y_pred']\n",
    "\n",
    "evaluate_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "796ff8c7",
   "metadata": {},
   "outputs": [
    {
     "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>y_test</th>\n",
       "      <th>y_pred</th>\n",
       "      <th>residuals</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>139482.000000</td>\n",
       "      <td>139482.000000</td>\n",
       "      <td>139482.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4.210855</td>\n",
       "      <td>4.206412</td>\n",
       "      <td>0.004443</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2.762196</td>\n",
       "      <td>1.446086</td>\n",
       "      <td>2.276826</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-8.832000</td>\n",
       "      <td>0.205833</td>\n",
       "      <td>-14.073562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2.555000</td>\n",
       "      <td>3.319029</td>\n",
       "      <td>-1.086534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>3.585000</td>\n",
       "      <td>3.851692</td>\n",
       "      <td>-0.124354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>5.528750</td>\n",
       "      <td>4.605443</td>\n",
       "      <td>0.891478</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>55.490000</td>\n",
       "      <td>58.489219</td>\n",
       "      <td>44.609461</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              y_test         y_pred      residuals\n",
       "count  139482.000000  139482.000000  139482.000000\n",
       "mean        4.210855       4.206412       0.004443\n",
       "std         2.762196       1.446086       2.276826\n",
       "min        -8.832000       0.205833     -14.073562\n",
       "25%         2.555000       3.319029      -1.086534\n",
       "50%         3.585000       3.851692      -0.124354\n",
       "75%         5.528750       4.605443       0.891478\n",
       "max        55.490000      58.489219      44.609461"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "evaluate_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ccaa09cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAFgCAYAAACFYaNMAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAABEo0lEQVR4nO3deZDk533f9/f3d/Z9zT17AwSwAAmBx5IURSWWRNE6rJi0y5IlWy7aUYWpimzLVU4cMnI5dlKOVEnFZVbFsU1LsmFbsaWiTJOWVJQpyKJ8UBSWAAgujr2APWbnnu7pu3/nkz+6dzm73AvEznTP7PdVtdXTv+4ZfHew+9lnnt/zfB8xxqCUUmrvWeMuQCmlHlYawEopNSYawEopNSYawEopNSYawEopNSbOuAt4UKanp83x48fHXYZSSt3k61//+qYxZuZ2rx2YAD5+/DinT58edxlKKXUTEbl8p9d0CkIppcZk7AEsIhUR+ZyIvC4ir4nIh0SkJiJfFpHzo8fquOtUSqkHbewBDHwG+JIx5iTwDPAa8CngOWPMY8Bzo+dKKXWgjDWARaQE/NfALwMYY0JjzDbwMeDZ0dueBT4+jvqUUmo3jXsE/AiwAfxTEXlRRH5JRPLAnDFmBWD0ODvOIpVSajeMO4Ad4L3APzTGvAfo8hamG0TkkyJyWkROb2xs7FaNSim1K8YdwEvAkjHma6Pnn2MYyGsisgAwely/3ScbYz5rjDlljDk1M3PbZXZKKTWxxhrAxphV4KqIPDG69BHgVeCLwCdG1z4BfGEM5Sml1K6ahI0YfwX4VRHxgDeAv8TwH4ZfF5GfAa4APz7G+pRSaleMPYCNMS8Bp27z0kf2uBSllNpT454DVkqph5YGsFJKjYkGsFJKjYkGsFJKjYkGsFJKjYkGsFJKjYkGsFJKjYkGsFJKjYkGsFJKjYkGsFJKjcnYtyIrpdR+YIyh0YvoBjF536GacxGRt/U1NYCVUuoejDGcXW2z1OjfuHa4muWJ+eLbCmGdglBKqXto9KKbwhdgqdGn0Yve1tfVAFZKqXvoBvFbun6/NICVUuoe8v7tZ2vvdP1+aQArpdQ9VHMuh6vZm64drmap5ty39XX1JpxSSt2DiPDEfJHZUkZXQSil1F4TEWp5j1ree2BfU6cglFJqTDSAlVJqTDSAlVJqTDSAlVJqTDSAlVJqTHQVxENsN5qLKKXunwbwQ2q3mosope6fTkE8pHaruYhS6v5pAD+kdqu5iFLq/mkAP6R2q7mIUur+aQA/pHaruYhS6v7pcOchtVvNRZRS908D+CG2G81FlFL3T6cglFJqTDSAlVJqTMY+BSEil4A2kACxMeaUiNSAXwOOA5eAnzDGNMZVo1JK7YZJGQF/vzHm3caYU6PnnwKeM8Y8Bjw3eq6UUgfKpATwrT4GPDv6+Fng4+MrRSmldsckBLAB/r2IfF1EPjm6NmeMWQEYPc7e7hNF5JMiclpETm9sbOxRuUop9WCMfQ4Y+LAxZllEZoEvi8jr9/uJxpjPAp8FOHXqlNmtApVSajeMfQRsjFkePa4Dnwc+AKyJyALA6HF9fBUqpdTuGGsAi0heRIrXPwb+OHAG+CLwidHbPgF8YTwVKqXU7hn3FMQc8PnR9lcH+P+MMV8SkeeBXxeRnwGuAD8+xhqVUmpXjDWAjTFvAM/c5voW8JG9r0gppfbO2OeAlVLqYaUBrJRSY6IBrJRSY6IBrJRSY6IBrJRSY6IBrJRSY6IBrJRSY6IBrJRSY6IBrJRSY6IBrJRSYzLuXhBqFxhjaPQiPW5eqQmnAXzAGGM4u9pmqdG/ce1wNcsT80UNYaUmjE5BHDCNXnRT+AIsNfo0etGYKlJK3YkG8AHTDeK3dF0pNT4awAdM3r/9rNKdriulxkcD+ICp5lwOV7M3XTtczVLNuWOqSCl1JzosOmBEhCfmi8yWMroKQqkJpwF8AIkItbxHLe+NuxSl1F3oFIRSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JroLYp7ThjlL7nwbwPqQNd5Q6GHQKYh/ShjtKHQwawPuQNtxR6mDQAN6HtOGOUgeDBvA+pA13lDoYdMi0D2nDHaUOBg3gfUob7ii1/+kUhFJKjYkGsFJKjcnYA1hEbBF5UUR+c/S8JiJfFpHzo8fquGtUSqndMPYABn4OeG3H808BzxljHgOeGz1XSqkDZ6wBLCKHgT8B/NKOyx8Dnh19/Czw8T0uSyml9sS4R8B/H/gbQLrj2pwxZgVg9Dh7p08WkU+KyGkROb2xsbGrhSql1IM2tgAWkR8D1o0xX/9Ov4Yx5rPGmFPGmFMzMzMPsDqllNp941wH/GHgT4rIjwIZoCQi/xJYE5EFY8yKiCwA62OsUSmlds3YRsDGmE8bYw4bY44DPwn8njHmp4EvAp8Yve0TwBfGVKJSSu2qcc8B384vAh8VkfPAR0fPlVLqwJmIrcjGmN8Hfn/08RbwkXHWo5RSe2ESR8BKKfVQ0ABWSqkx0QBWSqkx0QBWSqkx0QBWSqkx0QBWSqkx0QBWSqkx0QBWSqkx0QBWSqkxmYidcOrBMsbQ6EV6YrJSE04D+IAxxnB2tc1So3/j2uFqlifmixrCSk0YnYLYh4wx1LshV+s96t0QY8yN1xq96KbwBVhq9Gn0or0uUyl1DzoC3mfuNcLtBvFtP68bxNTy3l6VqZS6DzoC3mfuNcLN+7f/N/VO15VS46MBvM/cbYQLUM25HK5mb3rtcDVLNefuem1KqbdGh0X7zL1GuCLCE/NFZksZXQWh1ITTAN5nro9wb50D3jnCFRFqeU/nfJWacBrA+4yOcJU6ODSA9yEd4Sp1MOhNOKWUGhMNYKWUGhMNYKWUGhMNYKWUGhMNYKWUGhNdBXFAaAtKpfYfDeADQFtQKrU/6RTEAaAtKJXanzSAD4B7NehRSk0mDeADQFtQKrU/aQAfANqCUqn9SYdIB4A26FFqf9IAPiC0QY9S+48G8ENK1w0rNX5jDWARyQB/APijWj5njPlfRaQG/BpwHLgE/IQxpjGuOifZdxKkum5Yqckw7ptwAfADxphngHcDPywi3w18CnjOGPMY8NzoubrF9SB94XLjpsedx9Tfjq4bVmoyjDWAzVBn9NQd/TLAx4BnR9efBT6+99VNvu80SHXdsFKTYdwjYETEFpGXgHXgy8aYrwFzxpgVgNHj7B0+95MiclpETm9sbOxZzZOiO4hIUkMvjElSgzAc+d4rSHXdsFKTYewBbIxJjDHvBg4DHxCRd72Fz/2sMeaUMebUzMzMrtU4iYwxrLUDzlxrcm6tw5lrTerdCMHcM0h13bBSk2FihjzGmG0R+X3gh4E1EVkwxqyIyALD0bHaodGLaPRCpgoem50QA1yp93hsbuaeQarrhpWaDGMdAYvIjIhURh9ngR8EXge+CHxi9LZPAF8YS4ETrBvECMKhapZHZwscrmZ5ZLZAreDfV5BeXzd8pJajlvc0fJUag3GPgBeAZ0XEZviPwa8bY35TRL4K/LqI/AxwBfjxcRY5ia5PMwhC0Xco+DadIKEXJNS7oY5oldoHxhrAxpiXgffc5voW8JG9r2j/uD6Pu9ToYzAsN/rkPIdrjR5Ljb6u61VqHxj3CFh9h3bO4642+6QpOBYYhoG71OgzW8ro1mSlJtjYV0Go79z1edxixsW25Eb4XqfrepWabDoCnlBvZYuxrutVan+65whYRP7F/VxTD85b3WKs63qV2p/uZ4j0zp1PRisW3rc75Si48xbjO83p6rpepfanOwawiHwa+F+ArIi0rl8GQuCze1DbQ+tOc7edILrx+q0he69+wNp+UqnJc8cANsb8AvALIvILxphP72FND73bzd0aDFvtgK9f3iaIEnzX5uRcgZMLpW8L0lvDtpJ1OLfW0faTSk2Y+5mC+E0RyRtjuiLy08B7gc8YYy7vcm0PrZ1rfHdeO3OtyVo7vHGt1QuZK2Wo5r0bgZvzbNZbA65tD268r5JzafZC2LFKQpepKTV+9xPA/xB4RkSeAf4G8MvAPwf+2G4W9jC73Zzu6nbvpvAFw5VGn7OrLVzHptELEYQkNaw2+yxWs8gocC+ud6jkPOxbbrl2g1gDWKkxup91wLEZ3n7/GMOR72eA4u6WpW7t1WBZ1o7xq2GrE7LWGrDRGfCVsxtcG+2IC+KEjU5IJ0huvNtzbYI4+bb/hi5TU2q87ieA26Mbcn8B+K3RKghd37THZoo+R2s5BOhHKe0g5ngtR5oOO9hvjkLXd2wECKJvBW7Btzla02VqSk2a+xkC/VngzwH/rTFmVUSOAv/X7palblXLezx9qEQp67LeHlDLucwXM+Q8B2EYwkGUUPJtjtZyOM63/m09Us3x+FyB7X6sqyCUmiD3DOBR6P4G8Njo0ibw+V2tSn0bEeGJhRKz5SyrzT5X632uZ+zRWo4r9R6+a2MQnj5UGs4fh8lNYavH1is1We4ZwCLy3wGfBGrAo8Ah4B+h3cr23PUQreZcLJEbqyRqeZfH5maoFXzyoxHxreGrlJo89zMF8bPAB4CvARhjzovIbc9oU3tDRHh8rkDWs9nqBEwVfI5Us4iIHjev1D5yPwEcGGPCHTuuHODu556rXWWMuWljRb0b0Q8TZov+Xbcw6244pSbL/QTwV0Tk+pbkjwL/A/DvdrcsdTd36hWR3qFZTzeIqeZcHR0rNWHuZxnap4AN4JvAfw/8tjHm53e1KnVXd+oVYd2lXeWdQrvRix54fUqp+3M/AfxXjDH/xBjz48aYP2OM+Sci8nO7Xpm6ozttoJgp+ndsS7kztAVDkhp6Ycx6a3DHNpdKqd11P1MQnwA+c8u1v3iba2qP3K5XxOFq9sYys9u1pbwe2hYpjW7MlUYP17FuTFvoVIRSe+9u7Sh/iuEGjBMi8sUdLxWBrd0u7GF2r5tl9+r/e32H2/VRbzXnDkO7kuEbS03+04VNUgPljMNs0edao6eNeZQag7uNgP8LsAJMA//3jutt4OXdLOphdv00jHvdLLvTxoo0TXl5aZsr9T6+Y+NYcKiauxHY/XCLat7DtS2yrsXVep9y1tPGPEqNwd36AV8GLgMfutsXEJGvGmPu+h51/97qaRg7GWN44co2Xzm7gWHYfPJoLQf0buyMK+d9St1v3XgzQBAn2phHqTF4EKciZx7A11Ajd1rhcD8nHDd6ERfXOzcWaRvgSr1HnH7rFI2CbzNd+FaQD0NaG/MoNQ4PYtijt9AfoLdzwnE3iPFc+6ZrO0e41ZzLkWoOjMFzLDqDiCcXyzx9qKw34JQagwcxAlYP0P2ecGyMod4NuVrvUe+GGGPIezYYg+8I/SgBDAIcqWYxxrDU6DNT8Jgp+BR8h2NTBcIo4fx6V5eiKTUG99OM5y8Dv2qMadzpLQ+2pIfb/ZxwfLsbdYcqGcQY4sQQxSm+Y5FxbD74aI2sY/HilW0A2kFMHKfU8i4GwTD525UnsSalHoT7mYKYB54XkReAXwF+x9w8XPoLu1LZQ+xerSNvd6Pu9bUOedeinHVwZgrYFmQcm9mCz/n17vDrYoiihGvbfQqZEp59fYw8uduV73dViFL70T2nIIwxf5NhL+BfZrgB47yI/B8i8ujo9TO7WqH6Nre7IRdFCevtgG9ea3Fxo8v5tS7NfsxWNwCG4VvvRpxb73Bho8uLVxrUu9FoDDy525UnsSalHpT7mgMejXhXR79ioAp8TkT+z12sTd3B7W7Iua7FSnMwWn5mSIzhwnqLIDL0w4gwMVytd8m4FuWMg2NbN1ZI3G678k73swJjt0xiTUo9KPczB/xXGW5H3gR+CfifjDGRiFjAeYYnJas9dLutyLNFj422w2Y7IIgN9c4Ag8UfXtwkm3EIw4T2IKGQsXlsrkAl5xHEKUdq3/px/u2swNgtk1iTUg/K/fwpngb+9Ghjxg3GmFREfmx3ylJ382036jybK5sdmr2IQWzYbA/wHZsoSXBdm+4g5pHZPPUrTRYqWWaKPoJQBObL2Zu2MN+ux8Q41whPYk1KPSj3cybc37rLa6892HLU/RKRGyG03uxzZqWNYFht9tnoBMwUfBbLPq1+iOdYOCK8c7FMagwyuvF2a5DdzwqMvTaJNSn1oIz15zgROQL8c4YrLVLgs8aYz4hIDfg14DhwCfiJuyyDeygZYzi70uJyvU+7H/Li1QauJSyWfKIkZbsX4liw2Yk4OpXj6UNlcp7FO+aKWKPphtsF2SQe3jmJNSn1IIx7Ii0G/rox5gURKQJfF5EvM1xt8Zwx5hdF5FMMm8L/z2Osc+LUuyHfvNbiSr0HGLY6IZ0gZibvsdUJ2GyHZBwbwWABrV7MYjVLwXeo5T0dQSo1Aca6E84Ys2KMeWH0cRt4jeGpyx8Dnh297Vng42MpcIJttAOu1HsYIOMON12stwM6YUw3iJkqeFRyLkdrOcI4BjEMooQXr2xzdrWtO9+UmgATsxVZRI4D72F4+vKcMWYFhiEN3PYUZhH5pIicFpHTGxsbe1brJEiNudGEI4wNj83mKfg2Gdem2YtIjOE/nt/g4maXs2sdLCOY0UkY59baXKn3SNP027YzK6X2zrinIAAQkQLwG8BfM8a07vfHY2PMZ4HPApw6deqhSQ9jDDnPJuMIjX5M1rUIooSPnpxntdXj3UcrXN7s8MzhKlES866FKr04pFWP6YUJYZwShAkXCh6tQUwQp/iuzcm5AicXSjo9odQeGXsAi4jLMHx/1Rjzb0aX10RkwRizIiILwPr4Kpws12++Xa13qeZctnsRjiUcmimQGkM7CGh04dHZEq+uNIejZIGs5/DmZhfPdThWy+G5wn8+v4XYQnbUQa3VC5krZagV/LH+HpV6WIx1CkKGQ61fBl4zxvy9HS99keHmD0aPX9jr2iZVvRvyyrUW31hq8fzlBv0ooR8kHJvOsVjOYIngeTanL23RCVPC2NANUpYaXY7U8uQ9myhOSVPDeicgStIbX3ujE7LaGozxd6fUw2Xcc8AfZtjM5wdE5KXRrx8FfhH4qIicBz46eq4Y3nxbbQ9YbfWYLWZwbMG24NVrLVZbAe1BwrV6j2PTBcSkZF2bziDCsiw2OwOWm30u1Yd9IhwbXPtbfwSEOx9tr5R68MY6BWGM+U/cuZ3lR/aylv0iNQaTpmRch6+cWydJ4Uglw1TJ43itwFTeZbGS48y1bZ5crLDc6JH1HFabAU/MFdnuxTT7EZc2uzx9qMR2L77p+KKZok4/KLVXxj4HrN6a+VKGmVKW3zyzQpIOl6DNVjJc2erRC1K6QUQniJkvZ7EwFLMOC6UsUZqytN2jGyY8MpUn71vMFn1qhQyWQMZxOFbL6mYHpfaQBvA+U817VLIOs8Us3UHEfDnLVifg4kaXuVKMb9v0owRbDE8fLnO13mO64LHc7JNzHXzHphvGxC3DI1OGYs6hkHE5OV+kuscbNLTRunrYaQDvEzvD6kgtx1PzBXzH5uXlJpbA47MFWoOIq40ec0WffMbhhcsNrm0HHCr7HJ/OEcaGME5ZbwU8tVjildUWJ+eLOBYg8m3ht5sBuXMrdRAn+I7NsVqWJ3QZnHqIaADvAztPhTAYeoOYWiHDq8tNlrcHpMYwX8rwjpk8i5UMh6s5Ntt9sl4OV4SFapYkMRypZXl9tY3n2lxY7xHEMW9sdvlvnl6gO4humn7Y7ZModm6lvj4H3epHzJQyTOkyOPWQGPcqCHUfrp8KYTBca/S5ut3nhct1KjmH+bLHoUoGWwy2ZbHWGnB+rc3FjR5hkvLYfAFLhM1eSDdM6YUJS/Ue59baJClcrvf4xrUm7UF000643T6JYudWahie3nyl3mOjHTyQr6/UfqABPMGun3x8abNDkhq6QcxmJySKErb7IVfqfYwZjiZThBeuNOiFKRutAUdrOc6vttkexLQGIXPFDMvbPS7Xe1RyPgbDIEoZhDFpmnJurX1Tj4jdPoli51bqG7/f0XWlHhYawBPq+hTAC5cbXK73OXOtyVYnxCalFURkPJvtfoRJ4WgtC6mhlvPp9ENyGZevnNvgWjPk4nqHWj7D+fU2YZxytJZnqz3gUCWH5whPLZaJ0hTLkptGuLt9EsV8KcNM4eYVFzMFj/lS5oF8faX2Aw3gCbVzCqDg20wVPFaafbphwvJ2n6zrcGWry5VGn06QgAiNfkAln+HSVpcwgSRNCaKUNzY6LFZy9MKEle0ej84VeWK+wEIpw3TeZb6UxbMtBHNjhHv9JIqdHuRJFNW8x/tPVDlUyVDM2ByqZHj/iSpVXQanHiJ6E25C7fxRXxAOVbN4jtAPEy5udim78ANPzFHI2sRJyjeuNCjnvGFbeyN4tjBX8omTlEGUYllQybnkPYepgst03iXj2PTDmHY/4dXlFkdqeXLesC/EXpxEYSHU8j55f7g8zrrjnhylDiYN4Al164/6gmCLRcW3mC/5rDYHvLnZJEoNU3mPdx0qs7Q94HgtRynnUPRd3tzo4FgW00WX7zpUYbXRxXUdrjV7tINh6J5cKFHIWKQIvTC+KQJ38ySKRi9iaXuAbQk5b/h7XdoeMFvWzSDq4aEBPKGqOZdDlQyvr3UIogTfsVisZLi82SFKDK+vtEiMkPNsGt2Q1iAhTQ0XN7pkPZuN9oBS1iXv27znSJV2d4DrOjz3+ioFz6WeiZgpZgFDNe9TyfsUfJtumFDbg9/f3W7yaQCrh4XOAU8wAXKuTd6zSBPDaystzq11uVrvcaiWJ+dbeI7FYjnL+bUOINR7MWdXmpSzLjnPRkT44jeu8WYj4PTlOkeqeRIMqYEz17a5th1wtd4j71oIdz6a/kHT4+aV0gCeWDt/RPcch0v1Hte2B2Q8i2LWoTOIOD6VpzgK4a1uwFor4PJWhyNTRV5dbmFEOH2pQTdMWWr0AOEbS9sUPJcgSjgylYM05fx6m4trHQ5VMnt23Ptu3+RTaj/Q4caE2vkjehAnGAypMUSxIe85mNQgIsyWs9gWzBQ9OoOEjGezvN3HsS26wfAEjGLGxbGEjc6AvO8iFtiWRbcf8d4nZmh0AmaKPrOlzJ5tA9bj5pXSEfDE2vmjuO9Y1DshjU5AJ4jp9EOOTuUo+Q6CsNkMOD5VwLagmvWwrOFocrsbMpX36IUxiUk5MVXgqfkiT8yWcG1wbIvfe3WNzW6EBZxbbfPqcpN6J9iT8+Gu3+Q7UsvpSc3qoaQj4Al1/Uf0pUaffphQyDj0opitbkg3SukOQrphxMn5Chc3uizGKd/72DRRaoiCiOPTBV5ZaeF7Dt+8WmexnGGrG5Jg86VXlqlkfXKeTYpQ7wa0gpBLF/uIwEzR531HK5xcLGsoKrWLNIAn2GzRJzWGy1sp1ZyDLT7bvYiZgkezF5H3PZY223z/k7OstwIavYjNZo/3HKvx4lKdfih4ccrH33OUjVafy/U+XmwIE9joBMyVfBYrWTY7A/pxyoWNFhnX5Wqjj4gwU8owXdSdaUrtFg3gCbSzE1k7iHlzvY1jC1FiKGYcZvMur6+0sQSOTuU5t9rhaqNHOetyqJLl0maPrU7McnNAKeOS81ocr2WxRChlXUrZiPVWQLMfY1sBpayLpClxAmfWt0kNrLT6zJV8PvLknI6CldolGsATaOc25ChKEFJevNpmoz0gSQ3ufInvfnSKlUaffpTy6mqLYsZlqxeRGMg4FsemchgRrmx1ObcK0wWPfpiwngRgoJRxyXsWhyoZUiM4rs1So0tqhsvfwijl4kaH9x6r6bpcpXaJBvAE6gYxBkMnSAjihIubfa7WOyyWc5xf7/BHl+pUMjY/9PQi270QAdqDiFY/Iu87tIKYnOewtj3AEotGLyRjD2/MpRh6g4TZks9MwafeDTi30WGx7OM5DqkZNkevZl380UoKDWCldocG8ATKeRaXNjust0McC7a6AbPFLGvtgCQ1zBY8Fqs5Tl9u0AsiLm/1OFLLY9kWgzjlyYUCgzCiWvAoZhyOVX1cx8ZxhJLvkeQMOddhqztgoxPw9GKZ2VKG4maPo7Xs8L9RzFDOeboxQqldpH+7Jowxhjc3uzR7MavNPpYIxhgcy2K9FZD1bIpZl7VWn16YMlvyOVLN0Q0iZgo+GcdmvuARJYZmkBBFCcWsx++9ts72IGGu6FPvhbT6Ecem8uQ9m0ubXWwR5ssZHNuilveYLfgcqQ3PnQN0ja5Su0ADeMI0ehEX17vYlvDYXJGlepdHpnP0wpSca1HIOFSyDludiI12QCXrYlsWYWLoRyndICY1huPTudFKCYczS21SsYjSiCBJWWkG1PIejm3RDmJKGYecb+PawiMzJR6fL2KLUO+GbPeGG0Ie5HFESqkhDeAJ0w1iPMeiEyb0o5iV7QHV3DB0f+hdC2x0BoRRSjsIODlfIutZPH+pziBKAaEbxESp4cnFEsvNHu9arGBZQj9JcESYK2UIY0M564BJEREubnSYLmbohzFHporkHIvlVgg7eqMtNfrMljI6H6zUA6Q74SZM3rNp9ULSxNDuRWRci5VWyPNXmvzG168ylfdwLMOH3zFLsz8g49nMFDN89yNT5FyhlLEpZlyytvC9j05Ty/uEYcRT8wXetVjm0maXQRgTJwmHq3l6g5D3HKmy3QtJU/i911d5fa3DsLHwzR7UcURKqSEN4AljgJzvUszYzBR8jk3nWWn2CWKD5zp0BxGPz5e4tN7mxHSJjXaAZQmvrTQ5XM1T8FwcGc7/bnZDNlp9Ctnh/G2QJEwXPU4dr1Lybba6AadOTNMLE86utDi71qI9SLiw2qYz+Paw1RtySj1Y+jdqwvTChFrepZStsN4e0OqHWAZqORfPMmRcm8tbPVIsvrHUYL6cZbboUfZdGv2Q4zM5Mo7Nl19dxXcdcp5F0je0+hHlnI9gsbLdo5Dx2O6HnF1rc36tzTtmi7T7ERvtAc0gxLrln2btVKbUg6cBPGHyvoNgMCbBIqWctfmxZ+ZZaQ5YLFd54Uqdk/Ml4opHL0wIo4SM61IrOFhAwbMp+jYL5Sy9yFDwbLpBTGsQ0x4krLUDnpwv0gtjrtZ7PHOkygWg1Y/IeRa+O9zmLBjee7RCN0y0U5lSu0QDeMKUM8MTLr76xiabnZD11oCZcpYwijlayzNVyBImBs+x6UURg17KKytt5ko+1ZxDLe/SCVOynk03Cql3YxAh71lU8z7b/Yg3Nru873iNSs6n3h7ezKsVXATh5aVtjk9l6QTDY+OP1HLj/pYodWDpHPCEWWr0ObPSoj2IyfkOj86VsAykKXzl/AYXNjpcqffpDCLmy1lag4hy1qUbxCyUc4gIX7vUYKkxoN4OqBV8Gp2Ap49UqHcDjtXyvPdIlaMVD982zFczPDKdH7WfNHzPO2YoezYvL9VZb/XH/e1Q6kDTEfCEWdruU2/3KWV9fve1NYI4pZx1OD6dJxpEuLaN51o0uhGOJXz/E7PkPQvfsVnZHvDGZpcoSSj5LsenygzimO87Oct2d0CYGKw4JjEW1WyRqCyEaYptDIeKBeqDmDc2epiyRxAm1DvBuL8dSh1oOgKeMBbCdCnLi1fqCMOTKyxLeOnqNkem8lgCcWx4Yr5ArehzZauDiMXvvLLKVi9isx1SyXqstgacX+/QCVLWWgOiVIjihGrW4cmFMmdWO3zx5Wt87vQSL1xt0Y5TfEeYLvm8fK1FxnMIQ112ptRuGusIWER+BfgxYN0Y867RtRrwa8Bx4BLwE8aYxrhq3GtZNyXn2QDMlHyu1nvkPI9SJkc175GkhihJ6QwSLm128F2Xr17cxHUcUpMO543bIYvVDJ1BzEa7zzump9noDPi+J2bphAn/4ewGjV7IdCFLLZfSCSL6UcrJ+SKeO2Aq71D0bFp9DWCldtO4R8D/DPjhW659CnjOGPMY8Nzo+UOjPUgJooRKziHvWrzncIWjoxthy9t9rjR6PLVYBMtQyflUcy6zpSzdICJNh6Ft2SAIWdfm6cUqaRrh2Tbb/YgoSSlkXFabAW9udEgB2xaWtwc8+9XLfOnVdc5ca2HZ9nBRslJq14w1gI0xfwDUb7n8MeDZ0cfPAh/fy5rGrTOI6Ach7z1aw7Yt5koZlrd7LJazzORd3n+khC1CFMM3l7a5sN7l8laH6UIGEWj2Qg5Xsjwyk+NwLctyo43jevTjhLVWgCDUOwGLlSyPzRWIE0POtemEERnXphfE9GPDS0tNEob9IPbifDilHkbjHgHfzpwxZgVg9Dg75nr2jDGG9iBmrRPz1Te2yHo2WFDN+fT7IY7j0BgYtvsxf/jGJo/MFGn2Ao7U8rT7AYfKmeFJw0Wfte0+l7e6PDZfZaneYa7oY4whSWIWy1kWyhlqBZ9SxmG26NPuR3j28JDMhXKGjGVoRyEvXG5wdrWtIazULpjEAL5vIvJJETktIqc3NjbGXc7b1uhFrDUD1loBgzClH6TUOyHLjS6PLZb5zxc2ePlai/NrHU4ulEnShPcer3Kk6vGT7z/KTMFnKp+h14/wXZvH54qIpDx9qMp6s8dU3mW24LFQzZDzbOI0JYgTXMdiuphhqxsRxilvrHewHZeqO9z5ttTo0+hFY/7uKHXwTGIAr4nIAsDocf1ObzTGfNYYc8oYc2pmZmbPCtwt3SBGbEGA7X7McqvP62st/uvHZzm71kYQjlQz1Aouni0UMg5zBR/LctjsBlgWrLR6HJ0tUM7avLrc5JXlDqvNHicXKpSzDnnf44VLDV5ZbnF2pU3GtdloBVRzLludgCgxOI7Fdi8kHu7Ju1GbUurBmsR1wF8EPgH84ujxC+MtZ+/kfQfXAktgruxhiRDFKYMwoeg7HJ8p8Mq1Jo/NFOhFMcWMy3++WGe1FfDUQoFy1uXMUovXV9q861CJ2UIWWwyHa3ne3OpCalhphbi2xeFaljA25DyHrU7AoVqO9x+v4TjCarNPO0jY7IdYrkXBd7URj1K7YKwjYBH5V8BXgSdEZElEfoZh8H5URM4DHx09fyhUsg7FjMOjs8XhpoqNLs1+RK2YoR8mdAYxRsBzLBxbWGkOyPsOTx8qsbw94NJWjxPTeaLEsNwYUMm7VAo+v392k29cbbLRidloDUAgjlJyrk3Ws5kvZzAmJUhSzq60eXOzjzGGqaxHFKXaiEepXTLWYY0x5qfu8NJH9rSQCWCM4dxaB9+2uNbskfdtqjkP37VZqrd5arHM7762ytFanumiRzl1qXdbdIMYS2AQp+R9myhJqRU8DtdybHVCilmHC+sd5koZzq21eHKhSMZxiE3KK9daTBV9DlUzvGOmyEZjg5zvEsYJH35sGiHmkbmynoSh1C7RnysnxPWj6IMkpZJx2PQclht9KjmHrFcgSVM+cKJG3haKeZ9GLwRSEiMUsw7TkUcYxdQKHrYlbHUCbNsiTgy2JYhAOetxZCrPpfUu/SjhUDVLN0g4v9ql3Y/5nkdqbHT6zBRmiKOQTuyhix+U2j2TeBPuoXT9JpdvWyTGkHMtTs4XmC5kObPc4mqjx1Y3JEB4/nKdrCt8/8k5nj5UhBSO1rIcmcpzuJwh7wmPzxY4XPGZLrh84ESNjVZAaxBS8BzyGYec57DeGtALY5IkxRJhqdHnA0dmaPcjRGz6QUSjG+oKCKV2iY6AJ8T1PsBxKmAsfMdioxPRGoQ8tVAC4NJml04vZqaSYakRcGa5RWcQc2w6z8ZWn6NTeY5PF6jmXeLYcGGpQ5CkRLHhmSNlDlWzXNnqkvcdwijm5EKZ59+skxiD0+yzUM7QTyJcS8h6DlfrfWYrJbpBrGfBKbULdAQ8Iao5l6nCcMdasxdRcB1Ozhd5fKZIsxex1gpALPI5l5evNLAsi04Q0woSXllpYcTi5WstgjhlvR3yW2fWaPQiZgoZcr5D1rMJowTPtji/2ibB4g/f2OSxucKwgXspy8tXG9i2TTOIyXgOKRZxlOgKCKV2iQbwhBARagWfvGdjW4Zs1uHNjTbb/RDbtvAci0rWpd4ZcLiWYxDFpCn4jpB1bKbyHtWcSz+MGYTD1buLlRxZzyIIE5q9mI1exL9/bZ1mPwJSjtbyREnKe45WeXOzzWwpSxQnFDMu//HsGpe2uuQzwxOZlVIPngbwBCn4Lr0w4uRimbVGjycXijx9uMJ2N+TSZhdLDO9cLBPFCYM4ZaGSwXMswjihG0RMF3zyvkPGhbmiT6MfcnmrDyJYNtijG2r9OCWIU8I4wXdtwsQwV8pRyTr4DuRdi14M7z1aw7WFbe2KptSu0ACeINWcS9ZziaKIfNYHYxHFMaeOVfjIU3McKmcJ45D3HZ8atal0KPgO7z8xRc6zmS36fOnMMkerOY5N57lW77HVDQjjmMVShkdnc0wXPJIkZbMdcriaox/GWIAt8MhMgSSFC5sdDtdyLFaziFi6C06pXaI/W04QEeFYzWOtleeFKyvkPYtza20SA6SGR+fyHMnnaPQCFspZPMdi0c6x2uyPdrSFzJZyZDyH1e0eH37HFEen8viOxXpzeETRqeM16t2ArU5AzhWm54rUuwHvP1FhruhybavHVD4DlkXWG26+0DlgpXaH/s2aMK7tsrzd4ZljVTZaA6IEGr2QY1M53tjoU8l4RGmKYwntfsyZ5SaWJTw+V+TiRoeMa1HMuthisCyL3375GlnPY7M94P2PTLO61eboVI53zheJ05SXl1s4lk056xFEERe3+pRyHotFj4Jv6y44pXaRTkFMEGMMy80BG+2Ieieg0Y14fK5IxhGKvksxY3NiOk+SJGRci1LW4U+95xCPzeTIeRbVrMsPPjXPm+vbfO9jc/yXCxuUc1mWGj1iA9+4WudwNc+Z5Q6bnZArjQGrzZBGJ8S2BMdxuVTv4tkW3/PIFO87VtNdcErtIg3gCdLoRfSDlE6Ycrne49p2n7Vmj2eOVPEci6PVDNu9AbPlLAXPJk4N7X7Ae49NY4nFfCXLv/vGEmdWerSCkMO1PMWMTdZz8B2LXpgOO64BOd9hvRVgWzBd9OgHMUkCi+UcK9t9OkFMNedq+Cq1izSAJ0g3iOmGEVudgI32gEem81TyPq8sN9nqDChlPc5u9Hjh8jbHajm+61CZ6VKWr72xQZoamv2Y6UKGtVafnOey2R5gAaWMM9pSbBBLKPoWBd+mlHXJOhaea5GkCWKBkOK61nCEXO/dsxG7MYZ6N+RqvaenZyj1Fukc8ATJ+w5RCllPSIzPufUO7zpUZqGc5d1Hyry01GI651CpZglTw+trHabzLovVHOfX2tS7EbNFn3cdqrLR6vEDJ+c4t9riaDVLCixWsoRxwlSuMJwvdiwKWRcxhmIuQxhGIIJvW2x0BvzhxS36YXLHaQhjDGdX2yw1+jeuHa5mddpCqfukI+AJUs251LIuc8UsxhgKGRcBrm13GUQpzV7IfDkDRnh1tcWbm13CGK7WuxQzLlMFn1zG4ZVr24DNmaUG7z9R452LJap5j6v1LmutAM+xqOZcposuOc+mPRiOvB3X5bGZAnnfphemeK5919MwrjcQ2klPz1Dq/ukIeEIYY2j0IlxbeGQqz5V6jyPVHMenh4dnRnFCNecSJSkJKUE07HLW7IdYYhGnKdMFj2Yv5EgtR7MX0o0M51c7NIOYNze6pEA/jGn1I45PZTk2XeD15RbVgk+9E3J2pcOTCyUIEjxnOE0B3LEXxJ3WB2vvCKXujwbwBDDG8Ppyk9fXO/SjiHq/x/uPVYkMnFvrEYQR7z1e431Hq+Q9i61uzEtXGxR8D8casFDO4DlClMB2N2S7F1Er+Jg0xbItlhp9wiQl5zl0UlhtBXzgxBQvX9nm5KEKrgVr2wOKWZ/l7T6Nfkwp54ExIHLHdcBv9bpS6mY6BTEB6p2AP7iwyXOvrbPaDDA4XNzs8sWXlnnpaoPZSo6vnFunG8ZYtsWVepec52AwpAn0woTEQJQYfFdYrGRZa/Z5fL7EQtnHsQyDOCXr2VTyHhnXwrUMOd+h4DlUsy6Ha1m2On3myhmemMuzut1nsxPedR1wNedyuJq96ZquG1bq/ulQZQJcrvf45rUWxkAx47B0aTiPmgKL1Ty/99oaH3xkijc2u8yVfN4xMzz/zbGHh2c2+iHFjEOYJHzXoTIGuLjZJU6g1Q9477EaL1yuYzD0g4injlXpRQlvbnUoZVxe6/Q5MZXnPUcqXK538RybI7UcJ6Zyd72hJiI8MV9ktpShG8TkfUeXrin1FmgAT4BBlIIBxwLbEipZj26QMJVzKGccpJaj4NsYB3Ku0OiHhEmKMZD1HPrdgF6Q8Mb6MFD7YYIt4DrgWDbb3ZAfeXqRdj/Gdy08y3CtMeC7H5lmoz2gNYhxbIupgsfzlxukWESp8EM5755hKiLU8p7O+Sr1HdApiAmwWMmwUPTJuBZrzYB+FDKVd3jfiSnq3YD1dkC9G3FyocSFjS7TeZ9GN6LRi3jhyjYL5Qxr7QHFjMt6a4DnWNQKHp0g4dx6m04Y0+yFbPdC3tjocnGrj+04fO2NTcCi6Ln0gohOmBIlAIbpgn/jJpxSanfoCHgCHK3leP8jVf7tS8vUfOHEVJ7WIOGVay0sSzg+nafoC1e2A9IUBlHMM4criDU8igjAd4eHeFYyNq5jIQL1TkTec0kSOHOtyVY3Zqrg0RlE5Fybaj7DhY02U3kfxGIQxniOTSU3bAZfyOioVqndpCPgCSAi5D2brGvhey6NfswgTvE8m2LGIYpT8p5DMTN8Xsn7vLnV5lpjeHx8Ygz1bsDV+vC4IUuEME6JkpSFaoZD5QxrrQCR4YqLQZziezauLay3AkoZZ3izzrY4OpWlmnE5MpXTm2lK7TIdAU+AejfgpaUmG50BhyoZlrcHDKKUeicg49nkPRvPdbm43mG66PPGZpcohU4n4FozZa7o89hMkX6UsNbpU3BdyjmX2dDHEQvfFVzbQjDU8h7tYLhjToB3LpRYrPg4AludgE4QcerYFPMFX2+mKbXLdAQ8AS5t9Xhzvcljs6VR05yE7V7IoWqOXphQyfmcW2uS812MGU4t1DshWc9hdTvgSr3HTCGDCIixWO+EXKv36AQJFzc7bHVj3n+iwsn5Es1uwKOzBQ5XMrg2TBVcsrbFRifAsS1Wtge8fLVBOu5vilIPAR0BT4AwSpmt5Pjq+S3+/AeOMVtyOTGTJ4wSZktVajmXRs9hux+SdbNsdgIMUMq6+K5FP0xJMWz3Qlw7AwiJgWYvxLKEKEl5bLZCtx/x+GyB7X7AWntAP0qpZD0KWYcSFq8sd0gNeI5NN0ze9u/r+u4+XaKm1O3pCHgCLJZ9ihkXsSwMMe+YLvL8G1u8udXnxSsNDtdyrLX6NPsxYZywWMmSJoZuEJN1bWaKPqvNPr5r8/ShMkEU0x7E9KKYSs7jSr1Lmhq2g4T/eHGTL72yzvOXttloR7y81OTV5Q6+5+NawiBKqGSH88hvx/VGPS9cbtz0qN3SlPoWHQFPgHzGAQNhFOGJzR++Wcd1LEq+zYmpHK+tbvPHHp/jK+dWWd3u88h0jvmSTyeI6QUJh6s5ZooulbzH2naH73nHNM+/uUUp49IeRDxzpEoYp1xc79APU2IDGbHY7ARkPZvL9R6Ha1lqpRzbYcpiJct04e2tgLhTo57ZUkbXDCs1ogE8ATbaASuNLk8dqlAPYlaaA8pZl5zvcLnexbEsrBQ+8uQCF9bapKRYDFs1zJcz+I5gifDqtRaHKlmiMGKhkqUfpjgWZOzh5o13Lpa5uNmmkhv+bx/EKTnXIk4MjmXRC2PynkO9E7LRDph6GzfitFGPUvemUxAToN4N6YcJ6+2QzGg5WjnrcWGtg2NZtAcRg1R46crWsEXkVp+c6/L4XIGi7zJfzNAJI0pZj0Y3ZL0bMZsfTkssVLPkMw6rzZALay0enythjOFKo8/VrR6HKjmO1rIMoogoThERijmPa9uDt9VWUhv1KHVvGsATwCQxiQipScmLzZ/4rkVKGYdyzqMTRBQzLp1BSJBAdxAzVcqy3O5zcW14YsWFjS6vr3R5eanJ1e0+WILrCK5jYVLo9CPW2gMsx+bFKw2mCz5/7B1TfOjRKb55bZs4NRyq5umHIbNFH4B2ENMJvvMA1kY9St2bDkcmQBinHKlmWa13qQcB9c6AY7UsOd+m3glZbwU45QxxkvLGZpfWIKaSdZiZHm4XvrAR0QkSDIbLWwFhbJjKDftJrLX6HBsdTW8hxAau1ns8faRCexCwUMkxU/BZ2mpTKwyb6mx1h1ufT0xlOVLNfUfTELdr1FPJOroqQqkddAQ8Zmma8ma9T4aE/+rkPCvNiNXGgJJv8+RCmdikYIFrDedsU2NYaw84s9zm7FqbqWKGNzeGJ2Z0g4S859IPY/pRwonpPFl3uJEj59q4tkUSJzi2hZjRKHcQ0g0TSrkMcZqw0hxg0uH26K1O+LamIa436jlSG+6qO7fW0VURSu2gI+Axu1rvEYQx2WyGMysdNra7nJjNsz2IcMKYD56YYrMdMF/K8munLzOV8/jwo1OEccp0wacfRBybzvHqchvftYnShBMzRRwRWoOIguew0uzxzkNl2oOQR2dnWK53mS54HK74HK7kKWUEi5ST8yWEDtW8QzVnY5Ab0xB3GrXe71pfXRWh1Leb2AAWkR8GPgPYwC8ZY35xzCXtiuXtDhnP4oUrDao5n4UjVVwx9BN4Y7NLKTScudrAOWGxWM6Sz7icXRkeI3Rpq0spO0fesXjnQpHWICLrOVRyLpWMw0IpS73TI5/x+d1XV6kWMtTbdZ4+XMGYlKO1ApCw0gpBLFqDmNdWWuQzHpZ0OVIrUO8EnFvt3Kh356Gb9zqUc2c4twcRBoNwczjrqgj1MJvIKQgRsYF/APwI8BTwUyLy1HirevCMMaw3AzZaIZvdmOcvN1hpDji3NeAr57e4vNXnt765guXYnF9t8vhCiYvrHXzX5Wq9j2PbfOmby1TyPv04YbGcZaGcoT2IubDV5Y8ubTFXLvD6WptBPAzLa82Qy/UeG52I33hhidOXm7QGkLVtmmFKOefxb164SqOXkHUttjrhTTXvPHTzbody3roR42q9z3Kjj+HmKQddFaEeZhMZwMAHgAvGmDeMMSHwr4GPjbmmB67Ri7jaHDBI4eJ6mzc3e2R9l7OrLc6udZgp+fQjw+srHco5nzhOmSkN+wbX8j6NbkCKxdm1NlnXoZxz8V2bK5sdkhRcx2ap2UMQrk+11vIuZ1fbhImhkvPoRSmvr7bxPJdWL6Gc9ci4Ds1+RGrA8O3TCdfX+N5tre+t4exYDM+kC761xVlXRaiH3aQOPw4BV3c8XwI+eOubROSTwCcBjh49ujeVPUDd0U62Zj9iquhT73ZpDyKSBHzHIjUGSyA1w1+tQYyNEJPSCyLEEtI0pZL3SBNDreDxB2c3qBYyhHGCMZAk4NrDvhFb3RArA45tgTHESUrGtQkZ3gx0HCFKEhCDYwtJauA2AXx91Hq3tb63hrNBqOVdDlVz5H1HV0EoxeSOgG/3t/LbbpcbYz5rjDlljDk1MzOzB2U9WMOlWS5F38EC5soeOdcmMSnTBZcwTDk5X6DkC1MFn812nyNTOdqDENexCaOEpxbLxFHMBx+tkSQJj8wUKPgW9W5EwbPIeMJ3PzpNsx8wU/Qp+A5PL5a4Uu/xyEyBIEqYyru4luBY0A0SPnRimjRJOTFduOta3rut9b1dOBuE2VKGI7Uctfy9jztS6qCTSVwGJCIfAv62MeaHRs8/DWCM+YU7fc6pU6fM6dOn96jCB8MYwyvX6vzOS1dZHxheW2kjScKh6Ty9IKHdj3Fs4T3Ha9RbPTK+Q5QkWNgsbfeZKvisNfscrmXphxGe7WAMXNvuIWLhOhaHK1lKHgyMRasfk3UsPGd4w63eGeB7Dk/OF6h5CUttsG3BpML7TkzxAydnEZG7rnK40yqIe92gU+phISJfN8acuu1rExrADnAO+AhwDXge+HPGmFfu9Dn7MYBhdCNuu8vLl+u0QkM3iqm6NoLF6iCklHFwsejFETnHwQEGaQqWRRDHuI6FnQqWZdGNI7K2g0hKEIM37PGDlQq2ZbEdRXiWRdW26WHYHkRUMx45ElzPYa6aIcBmuuBzpJrFst7eD0jajlKpuwfwRM4BG2NiEfnLwO8wXIb2K3cL3/1MRJirFvhotTDs0VDv8Y2rdWID2dTBc23i2JDLZij4Np1BhBUbXltr0RzE5FybF6/U2R7EfPD4NEE87I72J59ZZLac5Stn13np8hZYFnGSMohTZvIuR6cKHJ0uIAjvPVbdlaVgemKyUnc3kQEMYIz5beC3x13HXhIRjtZy9MOEpUafUVsGDs9lmCn5nF3tkKYQWtCLoDNIaPUiFit5KnFCmAx3uR2qZJkpZajmXE7OF2n2Qq5uD0jS4SqIhUqGmVLmxppcXYur1HhMbAA/rG7XQ6Gac2n0Inphgm0JvuNwqOJTytpkbItuELPcGlDwXURguuDe+HH/5EKJuVKG1daAME64tNUj7zs3bYjQtbhKjYf+zZtAt/vR/fqyLsHQ6kdstSPWOwOmCh5bnYBqbniixVw5g2VZbPfjGysNagWfWsHHGEPGdb7txpiuxVVqPDSA94nro9Q4hcv1PvmMzVE/Ty3nYgGPzBRwHZtqzsUSue20wp1G13pjTKnx0ADeJ66vuT231h4tiBZmix6LZZ/tXsSFjS7FjMu1Rp+jtRw5z77t19EbY0pNDg3gfeL66DXr2fTDBM+1Kfg2JjX0guHNNxguO+uF8W13siilJosG8D5y6yoJgH6ccHQqSznnEcQp/iiYu2FCbcz1KqXuTgN4n7l1Hjc1hnNrbQShuON9urJBqcmnf0v3oZ3zuMaYm0bEoCsblNovNID3OV3ZoNT+pQF8AOjKBqX2p0ltR6mUUgeeBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JBrBSSo2JNuPZx4wxNHqRdkFTap/SAN6njDGcXW1/Wx/gJ+aLGsJK7RM6BbFPNXrRTeELsNTo0+hFY6pIKfVWaQDvU90gfkvXlVKTRwN4n7rTmW96FpxS+4cG8D5VzbkcrmZvuqZnwSm1v+hwaR+bKfoApMYwX8pQzXt6A06pfUQDeB+63QoIS4Sqngmn1L6iUxD7kK6AUOpg0ADeh3QFhFIHgwbwPqQrIJQ6GMYWwCLy4yLyioikInLqltc+LSIXROSsiPzQuGqcVLoCQqmDYZxDpjPAnwb+8c6LIvIU8JPAO4FF4HdF5HFjTLL3JU4mEeGJ+SKzpYz2gVBqHxtbABtjXgNuFxofA/61MSYA3hSRC8AHgK/ubYWTTUSo5T1quvJBqX1rEueADwFXdzxfGl37NiLySRE5LSKnNzY29qQ4pZR6UHZ1BCwivwvM3+alnzfGfOFOn3aba+Z2bzTGfBb4LMCpU6du+x6llJpUuxrAxpgf/A4+bQk4suP5YWD5wVSklFKTYxKnIL4I/KSI+CJyAngM+KMx16SUUg/cOJeh/SkRWQI+BPyWiPwOgDHmFeDXgVeBLwE/qysglFIH0ThXQXwe+PwdXvu7wN/d24qUUmpvTeIUhFJKPRQ0gJVSakzEmIOxektENoDLd3h5Gtjcw3LeqkmvDya/Rq3v7Zn0+mDya7xTfceMMTO3+4QDE8B3IyKnjTGn7v3O8Zj0+mDya9T63p5Jrw8mv8bvpD6dglBKqTHRAFZKqTF5WAL4s+Mu4B4mvT6Y/Bq1vrdn0uuDya/xLdf3UMwBK6XUJHpYRsBKKTVxNICVUmpMDnQA74djj0Tkh0c1XBCRT42rjh31/IqIrIvImR3XaiLyZRE5P3qsjrG+IyLyH0TktdH/25+bpBpFJCMifyQi3xjV93cmqb4dddoi8qKI/OaE1ndJRL4pIi+JyOlJq1FEKiLyORF5ffRn8UPfSX0HOoD51rFHf7Dz4i3HHv0w8P+KiL3XxY3+m/8A+BHgKeCnRrWN0z9j+D3Z6VPAc8aYx4DnRs/HJQb+ujHmSeC7gZ8dfc8mpcYA+AFjzDPAu4EfFpHvnqD6rvs54LUdzyetPoDvN8a8e8fa2kmq8TPAl4wxJ4FnGH4v33p9xpgD/wv4feDUjuefBj694/nvAB8aQ10fAn7nTnWN8ft1HDiz4/lZYGH08QJwdtw17qjtC8BHJ7FGIAe8AHxwkupj2GP7OeAHgN+cxP/HwCVg+pZrE1EjUALeZLSI4e3Ud9BHwHdy38cePSR13MucMWYFYPQ4O+Z6ABCR48B7gK8xQTWOfrx/CVgHvmyMmaj6gL8P/A0g3XFtkuqD4Sk4/15Evi4inxxdm5QaHwE2gH86msb5JRHJfyf1jfNU5Adit4892mWTUse+IyIF4DeAv2aMaU3SidBm2L/63SJSAT4vIu8ac0k3iMiPAevGmK+LyPeNuZy7+bAxZllEZoEvi8jr4y5oBwd4L/BXjDFfE5HP8B1Oh+z7ADb7+9ijSanjXtZEZMEYsyIiCwxHdmMjIi7D8P1VY8y/GV2eqBoBjDHbIvL7DOfUJ6W+DwN/UkR+FMgAJRH5lxNUHwDGmOXR47qIfJ7hyeiTUuMSsDT6yQbgcwwD+C3X97BOQUzKsUfPA4+JyAkR8RjeGPziGOq4ly8Cnxh9/AmG865jIcOh7i8Drxlj/t6OlyaiRhGZGY18EZEs8IPA65NSnzHm08aYw8aY4wz/vP2eMeanJ6U+ABHJi0jx+sfAH2d4Q30iajTGrAJXReSJ0aWPMDzB563XN86J9j2YLP9TDP+1CoA1br7h9fPARYYT5z8yxhp/FDg3quXnJ+B79q+AFSAafe9+BphieNPm/OixNsb6vpfhNM3LwEujXz86KTUC3wW8OKrvDPC3Rtcnor5bav0+vnUTbmLqYzjH+o3Rr1eu/72YsBrfDZwe/X/+t0D1O6lPtyIrpdSYPKxTEEopNXYawEopNSYawEopNSYawEopNSYawEopNSYawEopNSYawEo9ICLyF0Xk/xl3HWr/0ABW6h7G0apUPRw0gNWBIyL/+/VG7aPnf1dE/upt3vd9IvIHIvJ5EXlVRP6RiFij1zoi8r+JyNeAD4nIT48arb8kIv/4eiiLyF8SkXMi8hWGfRaUum8awOog+mVGe/JHgfqTwK/e4b0fAP468DTwKMMG/gB5hj2RPwhsAX+WYYeudwMJ8OdHDVf+DsPg/SjDpvpK3bd93w1NqVsZYy6JyJaIvAeYA140xmzd4e1/ZIx5A0BE/hXDXhOfYxiyvzF6z0eA9wHPj9peZhl2uvog8PvGmI3R5/8a8Pju/K7UQaQBrA6qXwL+IsNe0b9yl/fd2gzl+vOBGfb1hWHf5meNMZ/e+UYR+fhtPl+p+6ZTEOqg+jzDPrzvZ3jk1J18YNQO1GI4zfCfbvOe54A/M2oOfv1wyGMMT+L4PhGZGvUo/vEH+jtQB56OgNWBZIwJReQ/ANs7RrK381XgFxnOAf8Bw+C+9Wu9KiJ/k+ERORbDVp0/a4z5QxH526OvscLw/DddMaHum7ajVAfSKChfAH7cGHP+Du/5PuB/NMb82B6WptQNOgWhDpzRMfUXGB4RftvwVWoS6AhYHXgi8jTwL265HIyWmCk1NhrASik1JjoFoZRSY6IBrJRSY6IBrJRSY6IBrJRSY/L/A+c+x+3R/fAGAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(5, 5))\n",
    "\n",
    "sns.scatterplot(\n",
    "    data=evaluate_df,\n",
    "    y='y_test',\n",
    "    x='y_pred',\n",
    "    ax=ax,\n",
    "    alpha=.3\n",
    ")\n",
    "\n",
    "right = evaluate_df.iloc[:,:-1].max().max() + 3\n",
    "left = evaluate_df.iloc[:,:-1].min().min() - 3\n",
    "ax.set_xlim(left, right)\n",
    "ax.set_ylim(left, right)\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "dc830c7d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(5, 5))\n",
    "\n",
    "sns.scatterplot(\n",
    "    data=evaluate_df,\n",
    "    y='residuals',\n",
    "    x='y_pred',\n",
    "    ax=ax,\n",
    "    alpha=.3\n",
    ")\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b82ceeca",
   "metadata": {},
   "source": [
    "## Residential Land Use Feature importance based on Mean Decrease in Impurity (MDI)\n",
    "Source: https://scikit-learn.org/stable/auto_examples/inspection/plot_permutation_importance.html#sphx-glr-auto-examples-inspection-plot-permutation-importance-py\n",
    "\n",
    "Further reading: https://scikit-learn.org/stable/auto_examples/ensemble/plot_forest_importances.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "3012253b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "sanitsub_8D                0.005784\n",
       "bldgdepth                  0.005880\n",
       "longitude                  0.006395\n",
       "resarea                    0.007825\n",
       "bsmtcode_5.0               0.008638\n",
       "builtfar                   0.010887\n",
       "lotarea                    0.010962\n",
       "latitude                   0.011559\n",
       "ext_G                      0.012909\n",
       "bctcb2020_10255001006.0    0.014553\n",
       "ycoord                     0.014999\n",
       "CNSTRCT_YR                 0.017944\n",
       "bsmtcode_2.0               0.018327\n",
       "lotfront                   0.022155\n",
       "numbldgs                   0.028636\n",
       "yearbuilt                  0.031823\n",
       "healtharea                 0.031941\n",
       "numfloors                  0.056731\n",
       "HEIGHTROOF                 0.073924\n",
       "bsmtcode_1.0               0.402918\n",
       "dtype: float64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_names = rf[:-1].get_feature_names_out()\n",
    "\n",
    "mdi_importances = (\n",
    "    pd\n",
    "    .Series(rf[-1].feature_importances_, index=feature_names)\n",
    "    .sort_values(ascending=True)\n",
    "    .tail(20)\n",
    ")\n",
    "\n",
    "mdi_importances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "2d57f62a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8, 6))\n",
    "\n",
    "mdi_importances.plot.barh(ax=ax)\n",
    "\n",
    "plt.title(\"Random Forest Feature Importances (MDI)\")\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "101063f8",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}