Tensorflow - LendingClub Data - Naive Analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# imports\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "# Viz imports\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "# Config matplotlib\n",
    "%matplotlib inline\n",
    "plt.rcParams[\"patch.force_edgecolor\"] = True # in matplotlib, edge borders are turned off by default.\n",
    "sns.set_style(\"darkgrid\") # set a grey grid as a background\n",
    "\n",
    "# turn off warnings\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "# tensorflow and sklearn\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "# importing models and layers\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense,Dropout\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "\n",
    "# metrics\n",
    "from sklearn.metrics import classification_report, confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('TensorFlow_FILES/DATA/lending_club_loan_two.csv')\n",
    "df_info=pd.read_csv('TensorFlow_FILES/DATA/lending_club_info.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_info=pd.read_csv('TensorFlow_FILES/DATA/lending_club_info.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pretty printing has been turned OFF\n"
     ]
    }
   ],
   "source": [
    "%pprint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>LoanStatNew</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>loan_amnt</td>\n",
       "      <td>The listed amount of the loan applied for by t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>term</td>\n",
       "      <td>The number of payments on the loan. Values are...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>int_rate</td>\n",
       "      <td>Interest Rate on the loan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>installment</td>\n",
       "      <td>The monthly payment owed by the borrower if th...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>grade</td>\n",
       "      <td>LC assigned loan grade</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>sub_grade</td>\n",
       "      <td>LC assigned loan subgrade</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>emp_title</td>\n",
       "      <td>The job title supplied by the Borrower when ap...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>emp_length</td>\n",
       "      <td>Employment length in years. Possible values ar...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>home_ownership</td>\n",
       "      <td>The home ownership status provided by the borr...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>annual_inc</td>\n",
       "      <td>The self-reported annual income provided by th...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>verification_status</td>\n",
       "      <td>Indicates if income was verified by LC, not ve...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>issue_d</td>\n",
       "      <td>The month which the loan was funded</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>loan_status</td>\n",
       "      <td>Current status of the loan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>purpose</td>\n",
       "      <td>A category provided by the borrower for the lo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>title</td>\n",
       "      <td>The loan title provided by the borrower</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>zip_code</td>\n",
       "      <td>The first 3 numbers of the zip code provided b...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>addr_state</td>\n",
       "      <td>The state provided by the borrower in the loan...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>dti</td>\n",
       "      <td>A ratio calculated using the borrower’s total ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>earliest_cr_line</td>\n",
       "      <td>The month the borrower's earliest reported cre...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>open_acc</td>\n",
       "      <td>The number of open credit lines in the borrowe...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>pub_rec</td>\n",
       "      <td>Number of derogatory public records</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>revol_bal</td>\n",
       "      <td>Total credit revolving balance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>revol_util</td>\n",
       "      <td>Revolving line utilization rate, or the amount...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>total_acc</td>\n",
       "      <td>The total number of credit lines currently in ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>initial_list_status</td>\n",
       "      <td>The initial listing status of the loan. Possib...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>application_type</td>\n",
       "      <td>Indicates whether the loan is an individual ap...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>mort_acc</td>\n",
       "      <td>Number of mortgage accounts.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>pub_rec_bankruptcies</td>\n",
       "      <td>Number of public record bankruptcies</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             LoanStatNew                                        Description\n",
       "0              loan_amnt  The listed amount of the loan applied for by t...\n",
       "1                   term  The number of payments on the loan. Values are...\n",
       "2               int_rate                          Interest Rate on the loan\n",
       "3            installment  The monthly payment owed by the borrower if th...\n",
       "4                  grade                             LC assigned loan grade\n",
       "5              sub_grade                          LC assigned loan subgrade\n",
       "6              emp_title  The job title supplied by the Borrower when ap...\n",
       "7             emp_length  Employment length in years. Possible values ar...\n",
       "8         home_ownership  The home ownership status provided by the borr...\n",
       "9             annual_inc  The self-reported annual income provided by th...\n",
       "10   verification_status  Indicates if income was verified by LC, not ve...\n",
       "11               issue_d                The month which the loan was funded\n",
       "12           loan_status                         Current status of the loan\n",
       "13               purpose  A category provided by the borrower for the lo...\n",
       "14                 title            The loan title provided by the borrower\n",
       "15              zip_code  The first 3 numbers of the zip code provided b...\n",
       "16            addr_state  The state provided by the borrower in the loan...\n",
       "17                   dti  A ratio calculated using the borrower’s total ...\n",
       "18      earliest_cr_line  The month the borrower's earliest reported cre...\n",
       "19              open_acc  The number of open credit lines in the borrowe...\n",
       "20               pub_rec                Number of derogatory public records\n",
       "21             revol_bal                     Total credit revolving balance\n",
       "22            revol_util  Revolving line utilization rate, or the amount...\n",
       "23             total_acc  The total number of credit lines currently in ...\n",
       "24   initial_list_status  The initial listing status of the loan. Possib...\n",
       "25      application_type  Indicates whether the loan is an individual ap...\n",
       "26              mort_acc                       Number of mortgage accounts.\n",
       "27  pub_rec_bankruptcies               Number of public record bankruptcies"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.options.display.max_colwidth = 500"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24    The initial listing status of the loan. Possible values are – W, F\n",
       "Name: Description, dtype: object"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_info[df_info.LoanStatNew == 'initial_list_status']['Description']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25    Indicates whether the loan is an individual application or a joint application with two co-borrowers\n",
       "Name: Description, dtype: object"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_info[df_info.LoanStatNew == 'application_type']['Description']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['INDIVIDUAL', 'JOINT', 'DIRECT_PAY'], dtype=object)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['application_type'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>loan_amnt</th>\n",
       "      <th>term</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>installment</th>\n",
       "      <th>grade</th>\n",
       "      <th>sub_grade</th>\n",
       "      <th>emp_title</th>\n",
       "      <th>emp_length</th>\n",
       "      <th>home_ownership</th>\n",
       "      <th>annual_inc</th>\n",
       "      <th>...</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>revol_bal</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>initial_list_status</th>\n",
       "      <th>application_type</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>address</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>36 months</td>\n",
       "      <td>11.44</td>\n",
       "      <td>329.48</td>\n",
       "      <td>B</td>\n",
       "      <td>B4</td>\n",
       "      <td>Marketing</td>\n",
       "      <td>10+ years</td>\n",
       "      <td>RENT</td>\n",
       "      <td>117000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>16.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>36369.0</td>\n",
       "      <td>41.8</td>\n",
       "      <td>25.0</td>\n",
       "      <td>w</td>\n",
       "      <td>INDIVIDUAL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0174 Michelle Gateway\\nMendozaberg, OK 22690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8000.0</td>\n",
       "      <td>36 months</td>\n",
       "      <td>11.99</td>\n",
       "      <td>265.68</td>\n",
       "      <td>B</td>\n",
       "      <td>B5</td>\n",
       "      <td>Credit analyst</td>\n",
       "      <td>4 years</td>\n",
       "      <td>MORTGAGE</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>20131.0</td>\n",
       "      <td>53.3</td>\n",
       "      <td>27.0</td>\n",
       "      <td>f</td>\n",
       "      <td>INDIVIDUAL</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1076 Carney Fort Apt. 347\\nLoganmouth, SD 05113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15600.0</td>\n",
       "      <td>36 months</td>\n",
       "      <td>10.49</td>\n",
       "      <td>506.97</td>\n",
       "      <td>B</td>\n",
       "      <td>B3</td>\n",
       "      <td>Statistician</td>\n",
       "      <td>&lt; 1 year</td>\n",
       "      <td>RENT</td>\n",
       "      <td>43057.0</td>\n",
       "      <td>...</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11987.0</td>\n",
       "      <td>92.2</td>\n",
       "      <td>26.0</td>\n",
       "      <td>f</td>\n",
       "      <td>INDIVIDUAL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>87025 Mark Dale Apt. 269\\nNew Sabrina, WV 05113</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   loan_amnt        term  int_rate  installment grade sub_grade  \\\n",
       "0    10000.0   36 months     11.44       329.48     B        B4   \n",
       "1     8000.0   36 months     11.99       265.68     B        B5   \n",
       "2    15600.0   36 months     10.49       506.97     B        B3   \n",
       "\n",
       "         emp_title emp_length home_ownership  annual_inc  ... open_acc  \\\n",
       "0        Marketing  10+ years           RENT    117000.0  ...     16.0   \n",
       "1  Credit analyst     4 years       MORTGAGE     65000.0  ...     17.0   \n",
       "2     Statistician   < 1 year           RENT     43057.0  ...     13.0   \n",
       "\n",
       "  pub_rec revol_bal revol_util total_acc  initial_list_status  \\\n",
       "0     0.0   36369.0       41.8      25.0                    w   \n",
       "1     0.0   20131.0       53.3      27.0                    f   \n",
       "2     0.0   11987.0       92.2      26.0                    f   \n",
       "\n",
       "  application_type  mort_acc  pub_rec_bankruptcies  \\\n",
       "0       INDIVIDUAL       0.0                   0.0   \n",
       "1       INDIVIDUAL       3.0                   0.0   \n",
       "2       INDIVIDUAL       0.0                   0.0   \n",
       "\n",
       "                                           address  \n",
       "0     0174 Michelle Gateway\\nMendozaberg, OK 22690  \n",
       "1  1076 Carney Fort Apt. 347\\nLoganmouth, SD 05113  \n",
       "2  87025 Mark Dale Apt. 269\\nNew Sabrina, WV 05113  \n",
       "\n",
       "[3 rows x 27 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- get rid of some high NaN but low utility columns like title, emp_title\n",
    "- change some of the columns like \n",
    "- extrapolate fills for revol\n",
    "\n",
    "emp_title               22927 kill - too many null to be useful X\n",
    "emp_length              18301 fill\n",
    "title                    1755 kill - not sure how it would be useful X\n",
    "revol_util                276 small number - use mean?\n",
    "mort_acc                37795 mostly empty X\n",
    "pub_rec_bankruptcies      535 update na to 0 X"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EDA viz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Fully Paid', 'Charged Off'], dtype=object)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loan_status.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "77673"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.loan_status == 'Charged Off'].loan_status.count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot object at 0x10fc59610>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(df.loan_status)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    10000.0\n",
       "1     8000.0\n",
       "2    15600.0\n",
       "3     7200.0\n",
       "4    24375.0\n",
       "5    20000.0\n",
       "6    18000.0\n",
       "7    13000.0\n",
       "8    18900.0\n",
       "9    26300.0\n",
       "Name: loan_amnt, dtype: float64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.loan_amnt > 0]['loan_amnt'][:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.figure(figsize=(8,5))\n",
    "# sns.distplot(df[df.loan_amnt > 0]['loan_amnt'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['B', 'A', 'C', 'E', 'D', 'F', 'G'], dtype=object)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.grade.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "B    116018\n",
       "C    105987\n",
       "A     64187\n",
       "D     63524\n",
       "E     31488\n",
       "F     11772\n",
       "G      3054\n",
       "Name: grade, dtype: int64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.grade.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['B4', 'B5', 'B3', 'A2', 'C5', 'C3', 'A1', 'B2', 'C1', 'A5', 'E4',\n",
       "       'A4', 'A3', 'D1', 'C2', 'B1', 'D3', 'D5', 'D2', 'E1', 'E2', 'E5',\n",
       "       'F4', 'E3', 'D4', 'G1', 'F5', 'G2', 'C4', 'F1', 'F3', 'G5', 'G4',\n",
       "       'F2', 'G3'], dtype=object)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sub_grade.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>loan_amnt</th>\n",
       "      <th>term</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>installment</th>\n",
       "      <th>grade</th>\n",
       "      <th>sub_grade</th>\n",
       "      <th>emp_title</th>\n",
       "      <th>emp_length</th>\n",
       "      <th>home_ownership</th>\n",
       "      <th>annual_inc</th>\n",
       "      <th>...</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>revol_bal</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>initial_list_status</th>\n",
       "      <th>application_type</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>address</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>36 months</td>\n",
       "      <td>11.44</td>\n",
       "      <td>329.48</td>\n",
       "      <td>B</td>\n",
       "      <td>B4</td>\n",
       "      <td>Marketing</td>\n",
       "      <td>10+ years</td>\n",
       "      <td>RENT</td>\n",
       "      <td>117000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>16.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>36369.0</td>\n",
       "      <td>41.8</td>\n",
       "      <td>25.0</td>\n",
       "      <td>w</td>\n",
       "      <td>INDIVIDUAL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0174 Michelle Gateway\\nMendozaberg, OK 22690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8000.0</td>\n",
       "      <td>36 months</td>\n",
       "      <td>11.99</td>\n",
       "      <td>265.68</td>\n",
       "      <td>B</td>\n",
       "      <td>B5</td>\n",
       "      <td>Credit analyst</td>\n",
       "      <td>4 years</td>\n",
       "      <td>MORTGAGE</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>20131.0</td>\n",
       "      <td>53.3</td>\n",
       "      <td>27.0</td>\n",
       "      <td>f</td>\n",
       "      <td>INDIVIDUAL</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1076 Carney Fort Apt. 347\\nLoganmouth, SD 05113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15600.0</td>\n",
       "      <td>36 months</td>\n",
       "      <td>10.49</td>\n",
       "      <td>506.97</td>\n",
       "      <td>B</td>\n",
       "      <td>B3</td>\n",
       "      <td>Statistician</td>\n",
       "      <td>&lt; 1 year</td>\n",
       "      <td>RENT</td>\n",
       "      <td>43057.0</td>\n",
       "      <td>...</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11987.0</td>\n",
       "      <td>92.2</td>\n",
       "      <td>26.0</td>\n",
       "      <td>f</td>\n",
       "      <td>INDIVIDUAL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>87025 Mark Dale Apt. 269\\nNew Sabrina, WV 05113</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   loan_amnt        term  int_rate  installment grade sub_grade  \\\n",
       "0    10000.0   36 months     11.44       329.48     B        B4   \n",
       "1     8000.0   36 months     11.99       265.68     B        B5   \n",
       "2    15600.0   36 months     10.49       506.97     B        B3   \n",
       "\n",
       "         emp_title emp_length home_ownership  annual_inc  ... open_acc  \\\n",
       "0        Marketing  10+ years           RENT    117000.0  ...     16.0   \n",
       "1  Credit analyst     4 years       MORTGAGE     65000.0  ...     17.0   \n",
       "2     Statistician   < 1 year           RENT     43057.0  ...     13.0   \n",
       "\n",
       "  pub_rec revol_bal revol_util total_acc  initial_list_status  \\\n",
       "0     0.0   36369.0       41.8      25.0                    w   \n",
       "1     0.0   20131.0       53.3      27.0                    f   \n",
       "2     0.0   11987.0       92.2      26.0                    f   \n",
       "\n",
       "  application_type  mort_acc  pub_rec_bankruptcies  \\\n",
       "0       INDIVIDUAL       0.0                   0.0   \n",
       "1       INDIVIDUAL       3.0                   0.0   \n",
       "2       INDIVIDUAL       0.0                   0.0   \n",
       "\n",
       "                                           address  \n",
       "0     0174 Michelle Gateway\\nMendozaberg, OK 22690  \n",
       "1  1076 Carney Fort Apt. 347\\nLoganmouth, SD 05113  \n",
       "2  87025 Mark Dale Apt. 269\\nNew Sabrina, WV 05113  \n",
       "\n",
       "[3 rows x 27 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.grade != 0][:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot object at 0x157e04290>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "#sns.countplot(df[df.grade != 0]['grade'])\n",
    "sns.countplot(df.grade)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "B3    26655\n",
       "B4    25601\n",
       "C1    23662\n",
       "C2    22580\n",
       "B2    22495\n",
       "B5    22085\n",
       "C3    21221\n",
       "C4    20280\n",
       "B1    19182\n",
       "A5    18526\n",
       "C5    18244\n",
       "D1    15993\n",
       "A4    15789\n",
       "D2    13951\n",
       "D3    12223\n",
       "D4    11657\n",
       "A3    10576\n",
       "A1     9729\n",
       "D5     9700\n",
       "A2     9567\n",
       "E1     7917\n",
       "E2     7431\n",
       "E3     6207\n",
       "E4     5361\n",
       "E5     4572\n",
       "F1     3536\n",
       "F2     2766\n",
       "F3     2286\n",
       "F4     1787\n",
       "F5     1397\n",
       "G1     1058\n",
       "G2      754\n",
       "G3      552\n",
       "G4      374\n",
       "G5      316\n",
       "Name: sub_grade, dtype: int64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sub_grade.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot object at 0x14c9f6910>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "#sns.countplot(df[df.grade != 0]['grade'])\n",
    "sns.countplot(df.sub_grade)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### correlation analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>loan_amnt</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>14113.888089</td>\n",
       "      <td>8357.441341</td>\n",
       "      <td>500.00</td>\n",
       "      <td>8000.00</td>\n",
       "      <td>12000.00</td>\n",
       "      <td>20000.00</td>\n",
       "      <td>40000.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>int_rate</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>13.639400</td>\n",
       "      <td>4.472157</td>\n",
       "      <td>5.32</td>\n",
       "      <td>10.49</td>\n",
       "      <td>13.33</td>\n",
       "      <td>16.49</td>\n",
       "      <td>30.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>installment</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>431.849698</td>\n",
       "      <td>250.727790</td>\n",
       "      <td>16.08</td>\n",
       "      <td>250.33</td>\n",
       "      <td>375.43</td>\n",
       "      <td>567.30</td>\n",
       "      <td>1533.81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>annual_inc</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>74203.175798</td>\n",
       "      <td>61637.621158</td>\n",
       "      <td>0.00</td>\n",
       "      <td>45000.00</td>\n",
       "      <td>64000.00</td>\n",
       "      <td>90000.00</td>\n",
       "      <td>8706582.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dti</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>17.379514</td>\n",
       "      <td>18.019092</td>\n",
       "      <td>0.00</td>\n",
       "      <td>11.28</td>\n",
       "      <td>16.91</td>\n",
       "      <td>22.98</td>\n",
       "      <td>9999.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>open_acc</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>11.311153</td>\n",
       "      <td>5.137649</td>\n",
       "      <td>0.00</td>\n",
       "      <td>8.00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>14.00</td>\n",
       "      <td>90.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pub_rec</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>0.178191</td>\n",
       "      <td>0.530671</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>86.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>revol_bal</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>15844.539853</td>\n",
       "      <td>20591.836109</td>\n",
       "      <td>0.00</td>\n",
       "      <td>6025.00</td>\n",
       "      <td>11181.00</td>\n",
       "      <td>19620.00</td>\n",
       "      <td>1743266.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>revol_util</th>\n",
       "      <td>395754.0</td>\n",
       "      <td>53.791749</td>\n",
       "      <td>24.452193</td>\n",
       "      <td>0.00</td>\n",
       "      <td>35.80</td>\n",
       "      <td>54.80</td>\n",
       "      <td>72.90</td>\n",
       "      <td>892.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>total_acc</th>\n",
       "      <td>396030.0</td>\n",
       "      <td>25.414744</td>\n",
       "      <td>11.886991</td>\n",
       "      <td>2.00</td>\n",
       "      <td>17.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>32.00</td>\n",
       "      <td>151.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mort_acc</th>\n",
       "      <td>358235.0</td>\n",
       "      <td>1.813991</td>\n",
       "      <td>2.147930</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>34.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <td>395495.0</td>\n",
       "      <td>0.121648</td>\n",
       "      <td>0.356174</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>8.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         count          mean           std     min       25%  \\\n",
       "loan_amnt             396030.0  14113.888089   8357.441341  500.00   8000.00   \n",
       "int_rate              396030.0     13.639400      4.472157    5.32     10.49   \n",
       "installment           396030.0    431.849698    250.727790   16.08    250.33   \n",
       "annual_inc            396030.0  74203.175798  61637.621158    0.00  45000.00   \n",
       "dti                   396030.0     17.379514     18.019092    0.00     11.28   \n",
       "open_acc              396030.0     11.311153      5.137649    0.00      8.00   \n",
       "pub_rec               396030.0      0.178191      0.530671    0.00      0.00   \n",
       "revol_bal             396030.0  15844.539853  20591.836109    0.00   6025.00   \n",
       "revol_util            395754.0     53.791749     24.452193    0.00     35.80   \n",
       "total_acc             396030.0     25.414744     11.886991    2.00     17.00   \n",
       "mort_acc              358235.0      1.813991      2.147930    0.00      0.00   \n",
       "pub_rec_bankruptcies  395495.0      0.121648      0.356174    0.00      0.00   \n",
       "\n",
       "                           50%       75%         max  \n",
       "loan_amnt             12000.00  20000.00    40000.00  \n",
       "int_rate                 13.33     16.49       30.99  \n",
       "installment             375.43    567.30     1533.81  \n",
       "annual_inc            64000.00  90000.00  8706582.00  \n",
       "dti                      16.91     22.98     9999.00  \n",
       "open_acc                 10.00     14.00       90.00  \n",
       "pub_rec                   0.00      0.00       86.00  \n",
       "revol_bal             11181.00  19620.00  1743266.00  \n",
       "revol_util               54.80     72.90      892.30  \n",
       "total_acc                24.00     32.00      151.00  \n",
       "mort_acc                  1.00      3.00       34.00  \n",
       "pub_rec_bankruptcies      0.00      0.00        8.00  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe().transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot object at 0x157b53810>"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(df.corr())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "loan_amnt                   0\n",
       "term                        0\n",
       "int_rate                    0\n",
       "installment                 0\n",
       "grade                       0\n",
       "sub_grade                   0\n",
       "emp_title               22927\n",
       "emp_length              18301\n",
       "home_ownership              0\n",
       "annual_inc                  0\n",
       "verification_status         0\n",
       "issue_d                     0\n",
       "loan_status                 0\n",
       "purpose                     0\n",
       "title                    1755\n",
       "dti                         0\n",
       "earliest_cr_line            0\n",
       "open_acc                    0\n",
       "pub_rec                     0\n",
       "revol_bal                   0\n",
       "revol_util                276\n",
       "total_acc                   0\n",
       "initial_list_status         0\n",
       "application_type            0\n",
       "mort_acc                37795\n",
       "pub_rec_bankruptcies      535\n",
       "address                     0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get a count of nulls per column - need to scrub out categories before I can use fillna()\n",
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Drops\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# consolidated drops - this tanks ram and kills the kernal\n",
    "# df.drop(['title','purpose','emp_title'],axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(['title'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(['purpose'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "df =  df.drop(['emp_title'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(['initial_list_status'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# if null, no bankruptcies\n",
    "df[df.pub_rec_bankruptcies.isnull()] = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### fixing columns that have numbers and nulls"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "loan_amnt                   0\n",
       "term                        0\n",
       "int_rate                    0\n",
       "installment                 0\n",
       "grade                       0\n",
       "sub_grade                   0\n",
       "emp_length              18301\n",
       "home_ownership              0\n",
       "annual_inc                  0\n",
       "verification_status         0\n",
       "issue_d                     0\n",
       "loan_status                 0\n",
       "dti                         0\n",
       "earliest_cr_line            0\n",
       "open_acc                    0\n",
       "pub_rec                     0\n",
       "revol_bal                   0\n",
       "revol_util                276\n",
       "total_acc                   0\n",
       "application_type            0\n",
       "mort_acc                37260\n",
       "pub_rec_bankruptcies        0\n",
       "address                     0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.revol_bal = df.revol_bal.apply(lambda x: x if isinstance(x,float) else x if isinstance(x,int) else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.mort_acc = df.mort_acc.apply(lambda x: x if isinstance(x,float) else x if isinstance(x,int) else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.mort_acc = df.mort_acc.fillna(0)\n",
    "df.mort_acc = df.mort_acc.apply(lambda x: x if isinstance(x,float) else x if isinstance(x,int) else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.revol_util = df.revol_util.fillna(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### get a list of columns that have NaNs\n",
    "\n",
    "everything left should be numbers that can have zeroes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['emp_length']"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nan_values = df.isna()\n",
    "nan_columns = nan_values.any()\n",
    "columns_with_nan = df.columns[nan_columns].tolist()\n",
    "columns_with_nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "df[columns_with_nan] = df[columns_with_nan].fillna(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### turn employment length into a number"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "def employ(x):\n",
    "    if x is not None:\n",
    "        if isinstance(x,float) or isinstance(x,int):\n",
    "            return x\n",
    "        elif '10+' in x:\n",
    "            return 10\n",
    "        elif '<' in x:\n",
    "            return 1\n",
    "        else:\n",
    "            return x\n",
    "    else:\n",
    "        return 0\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.emp_length = df.emp_length.apply(employ)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# first break out the string\n",
    "df.emp_length = df.emp_length.apply(lambda x: x.split(' ')[0] if isinstance(x, str) else x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df.emp_length.apply(lambda x: x.split('+')[0] if '+' in x else x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### reset term to just be a number"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     36 months\n",
       "1     36 months\n",
       "2     36 months\n",
       "Name: term, dtype: object"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.term[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'36'"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['term'].iloc[10].split(' ')[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "def term(x):\n",
    "    if x != 0:\n",
    "        return int(x.split(' ')[1])\n",
    "    else: \n",
    "        return int(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.term = df.term.apply(term)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fixing Categores with Dummies"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### using drop first with dummies\n",
    "not duplicating information - want to change Categories to either numeric or boolean. \n",
    "\n",
    "\n",
    "when encoding M/F you don't want to end up with M and F, you want it to be a Male Boolean M 0 or 1\n",
    "\n",
    "if A/B/C you don't want A, B and C columns - Boolean A or B and if not, then C\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>0</th>\n",
       "      <th>A</th>\n",
       "      <th>B</th>\n",
       "      <th>C</th>\n",
       "      <th>D</th>\n",
       "      <th>E</th>\n",
       "      <th>F</th>\n",
       "      <th>G</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  A  B  C  D  E  F  G\n",
       "0  0  0  1  0  0  0  0  0\n",
       "1  0  0  1  0  0  0  0  0\n",
       "2  0  0  1  0  0  0  0  0"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.get_dummies(df['grade'])[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "grade_dummies = pd.get_dummies(df['grade'],drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "df  = pd.concat([df.drop('grade',axis=1),grade_dummies],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "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>0</th>\n",
       "      <th>A1</th>\n",
       "      <th>A2</th>\n",
       "      <th>A3</th>\n",
       "      <th>A4</th>\n",
       "      <th>A5</th>\n",
       "      <th>B1</th>\n",
       "      <th>B2</th>\n",
       "      <th>B3</th>\n",
       "      <th>B4</th>\n",
       "      <th>...</th>\n",
       "      <th>F1</th>\n",
       "      <th>F2</th>\n",
       "      <th>F3</th>\n",
       "      <th>F4</th>\n",
       "      <th>F5</th>\n",
       "      <th>G1</th>\n",
       "      <th>G2</th>\n",
       "      <th>G3</th>\n",
       "      <th>G4</th>\n",
       "      <th>G5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  A1  A2  A3  A4  A5  B1  B2  B3  B4  ...  F1  F2  F3  F4  F5  G1  G2  G3  \\\n",
       "0  0   0   0   0   0   0   0   0   0   1  ...   0   0   0   0   0   0   0   0   \n",
       "1  0   0   0   0   0   0   0   0   0   0  ...   0   0   0   0   0   0   0   0   \n",
       "2  0   0   0   0   0   0   0   0   1   0  ...   0   0   0   0   0   0   0   0   \n",
       "\n",
       "   G4  G5  \n",
       "0   0   0  \n",
       "1   0   0  \n",
       "2   0   0  \n",
       "\n",
       "[3 rows x 36 columns]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.get_dummies(df['sub_grade'])[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "sub_grade_dummies = pd.get_dummies(df['sub_grade'],drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "df  = pd.concat([df.drop('sub_grade',axis=1),sub_grade_dummies],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "535"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# how many loan amounts are zero?\n",
    "df[df.loan_amnt < 1]['loan_amnt'].count() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "395495"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loan_amnt.count() - df[df.loan_amnt < 1]['loan_amnt'].count() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "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>loan_amnt</th>\n",
       "      <th>term</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>installment</th>\n",
       "      <th>emp_length</th>\n",
       "      <th>home_ownership</th>\n",
       "      <th>annual_inc</th>\n",
       "      <th>verification_status</th>\n",
       "      <th>issue_d</th>\n",
       "      <th>loan_status</th>\n",
       "      <th>...</th>\n",
       "      <th>F1</th>\n",
       "      <th>F2</th>\n",
       "      <th>F3</th>\n",
       "      <th>F4</th>\n",
       "      <th>F5</th>\n",
       "      <th>G1</th>\n",
       "      <th>G2</th>\n",
       "      <th>G3</th>\n",
       "      <th>G4</th>\n",
       "      <th>G5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>36</td>\n",
       "      <td>11.44</td>\n",
       "      <td>329.48</td>\n",
       "      <td>10</td>\n",
       "      <td>RENT</td>\n",
       "      <td>117000.0</td>\n",
       "      <td>Not Verified</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>Fully Paid</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8000.0</td>\n",
       "      <td>36</td>\n",
       "      <td>11.99</td>\n",
       "      <td>265.68</td>\n",
       "      <td>4</td>\n",
       "      <td>MORTGAGE</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>Not Verified</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>Fully Paid</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15600.0</td>\n",
       "      <td>36</td>\n",
       "      <td>10.49</td>\n",
       "      <td>506.97</td>\n",
       "      <td>1</td>\n",
       "      <td>RENT</td>\n",
       "      <td>43057.0</td>\n",
       "      <td>Source Verified</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>Fully Paid</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 63 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   loan_amnt  term  int_rate  installment emp_length home_ownership  \\\n",
       "0    10000.0    36     11.44       329.48         10           RENT   \n",
       "1     8000.0    36     11.99       265.68          4       MORTGAGE   \n",
       "2    15600.0    36     10.49       506.97          1           RENT   \n",
       "\n",
       "   annual_inc verification_status   issue_d loan_status  ...  F1 F2  F3  F4  \\\n",
       "0    117000.0        Not Verified  Jan-2015  Fully Paid  ...   0  0   0   0   \n",
       "1     65000.0        Not Verified  Jan-2015  Fully Paid  ...   0  0   0   0   \n",
       "2     43057.0     Source Verified  Jan-2015  Fully Paid  ...   0  0   0   0   \n",
       "\n",
       "   F5  G1  G2 G3  G4  G5  \n",
       "0   0   0   0  0   0   0  \n",
       "1   0   0   0  0   0   0  \n",
       "2   0   0   0  0   0   0  \n",
       "\n",
       "[3 rows x 63 columns]"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# what categories are left?\n",
    "# drop all the loan amount == 0 \n",
    "df[df.loan_amnt > 0].head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Verified           139562\n",
       "Source Verified    131385\n",
       "Not Verified       124548\n",
       "0                     535\n",
       "Name: verification_status, dtype: int64"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.verification_status.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "verification_dummies = pd.get_dummies(df['verification_status'],drop_first=True)\n",
    "df  = pd.concat([df.drop('verification_status',axis=1),verification_dummies],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dummies for home_ownership\n",
    "ownership_dummies = pd.get_dummies(df['home_ownership'],drop_first=True)\n",
    "df  = pd.concat([df.drop('home_ownership',axis=1),ownership_dummies],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dummies for application  status\n",
    "application_dummies = pd.get_dummies(df['application_type'],drop_first=True)\n",
    "df  = pd.concat([df.drop('application_type',axis=1),application_dummies],axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### conver the the primary value (loan status) to boolean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Fully Paid', 'Charged Off', 0], dtype=object)"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# convert to \"default\" boolean\n",
    "df.loan_status.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Fully Paid     317909\n",
       "Charged Off     77586\n",
       "0                 535\n",
       "Name: loan_status, dtype: int64"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loan_status.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dropping everything with status of 0\n",
    "df = df.drop(df[df.loan_status == 0].index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Now boolean for \"defaulted on loan\" to be true\n",
    "def statusbool(x):\n",
    "    if x == 'Charged Off':\n",
    "        return True\n",
    "    else:\n",
    "        return False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "def statusnum(x):\n",
    "    if x == 'Charged Off':\n",
    "        return 1\n",
    "    else:\n",
    "        return 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.loan_status = df.loan_status.apply(statusnum)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "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>loan_amnt</th>\n",
       "      <th>term</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>installment</th>\n",
       "      <th>emp_length</th>\n",
       "      <th>annual_inc</th>\n",
       "      <th>issue_d</th>\n",
       "      <th>loan_status</th>\n",
       "      <th>dti</th>\n",
       "      <th>earliest_cr_line</th>\n",
       "      <th>...</th>\n",
       "      <th>Verified</th>\n",
       "      <th>ANY</th>\n",
       "      <th>MORTGAGE</th>\n",
       "      <th>NONE</th>\n",
       "      <th>OTHER</th>\n",
       "      <th>OWN</th>\n",
       "      <th>RENT</th>\n",
       "      <th>DIRECT_PAY</th>\n",
       "      <th>INDIVIDUAL</th>\n",
       "      <th>JOINT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>36</td>\n",
       "      <td>11.44</td>\n",
       "      <td>329.48</td>\n",
       "      <td>10</td>\n",
       "      <td>117000.0</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>0</td>\n",
       "      <td>26.24</td>\n",
       "      <td>Jun-1990</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8000.0</td>\n",
       "      <td>36</td>\n",
       "      <td>11.99</td>\n",
       "      <td>265.68</td>\n",
       "      <td>4</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>0</td>\n",
       "      <td>22.05</td>\n",
       "      <td>Jul-2004</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15600.0</td>\n",
       "      <td>36</td>\n",
       "      <td>10.49</td>\n",
       "      <td>506.97</td>\n",
       "      <td>1</td>\n",
       "      <td>43057.0</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>0</td>\n",
       "      <td>12.79</td>\n",
       "      <td>Aug-2007</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 72 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   loan_amnt  term  int_rate  installment emp_length  annual_inc   issue_d  \\\n",
       "0    10000.0    36     11.44       329.48         10    117000.0  Jan-2015   \n",
       "1     8000.0    36     11.99       265.68          4     65000.0  Jan-2015   \n",
       "2    15600.0    36     10.49       506.97          1     43057.0  Jan-2015   \n",
       "\n",
       "   loan_status    dti earliest_cr_line  ...  Verified  ANY  MORTGAGE  NONE  \\\n",
       "0            0  26.24         Jun-1990  ...         0    0         0     0   \n",
       "1            0  22.05         Jul-2004  ...         0    0         1     0   \n",
       "2            0  12.79         Aug-2007  ...         0    0         0     0   \n",
       "\n",
       "   OTHER  OWN  RENT DIRECT_PAY  INDIVIDUAL  JOINT  \n",
       "0      0    0     1          0           1      0  \n",
       "1      0    0     0          0           1      0  \n",
       "2      0    0     1          0           1      0  \n",
       "\n",
       "[3 rows x 72 columns]"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[:3]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### convert issue date and the date of the earliest credit line into numbers\n",
    "\n",
    "Then subtract them to make a usable \"length between first credit and loan\" date which could be meaningful to default"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Jan', '2015']"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.issue_d !=0]['issue_d'].iloc[0].split('-')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "# map the dates to months\n",
    "months ={'Jan':'01','Feb':'02','Mar':'03','Apr':'04','May':'05','Jun':'06','Jul':'07','Aug':'08','Sep':'09','Oct':10,'Nov':11,'Dec':12}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "# fix issue_date\n",
    "def issue(x):\n",
    "    if isinstance(x,str):\n",
    "        month, year = x.split('-')[0],x.split('-')[1]\n",
    "        return year + '-' + str(months[month])\n",
    "    else: \n",
    "        return 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert these into numerics for further processing\n",
    "df.issue_d = df.issue_d.apply(issue)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.earliest_cr_line = df.earliest_cr_line.apply(issue)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to yearmonth integer\n",
    "# what does this indicate? Not months. not days.\n",
    "def timedelta(x):\n",
    "    if x !=0 and x is not None:\n",
    "        year, month = x.split('-')[0],x.split('-')[1]\n",
    "        mytime = year + month\n",
    "        return int(mytime)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "# now apply the timedelta to them both\n",
    "df.issue_d = df.issue_d.apply(timedelta)\n",
    "df.earliest_cr_line = df.earliest_cr_line.apply(timedelta)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "# now create a new feature to reflect the delta between the first credit line and the issue date\n",
    "df['acc_length'] = df.issue_d - df.earliest_cr_line"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    2495\n",
       "1    1094\n",
       "2     793\n",
       "Name: acc_length, dtype: int64"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['acc_length'][:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "# now drop the columns this after the issue/credit metric established\n",
    "# compound drop not working\n",
    "# df.drop([['issue_d','earliest_cr_line']], axis=1) #<- borked\n",
    "df = df.drop('issue_d',axis=1)\n",
    "df = df.drop('earliest_cr_line',axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**does zip code matter?**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'22690'"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.address.iloc[0].split(' ')[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "# testing the .apply() output\n",
    "# df.address.apply(lambda x: x.split()[-1] if isinstance(x,str) else x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['zipcode'] =  df.address.apply(lambda x: x.split()[-1] if isinstance(x,str) else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    22690\n",
       "1    05113\n",
       "2    05113\n",
       "Name: zipcode, dtype: object"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['zipcode'][:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Zip code is in place so get rid of address\n",
    "df.drop('address',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['22690', '05113', '00813', '11650', '30723', '70466', '29597',\n",
       "       '48052', '86630', '93700'], dtype=object)"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# not too many in this dataset, so lets add dummies\n",
    "df.zipcode.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dummies for zip code\n",
    "# dummies for home_ownership\n",
    "zipcode_dummies = pd.get_dummies(df['zipcode'],drop_first=True)\n",
    "df  = pd.concat([df.drop('zipcode',axis=1),ownership_dummies],axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### dealing with columns with only zeroes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([], dtype='object')"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find columns that have only zeros \n",
    "df.columns[(df == 0).all()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: []\n",
       "Index: [0, 1, 2]"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get all the columns data with:\n",
    "df.loc[:, (df == 0).all()][:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "loan_amnt    False\n",
       "term         False\n",
       "int_rate     False\n",
       "dtype: bool"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get the set of columns in a boolean:\n",
    "(df == 0).all()[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "# boolean everything without a zero\n",
    "#df[df == 0][:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "loan_amnt               535\n",
       "term                    535\n",
       "int_rate                535\n",
       "installment             535\n",
       "emp_length              535\n",
       "annual_inc              535\n",
       "loan_status             535\n",
       "dti                     535\n",
       "open_acc                535\n",
       "pub_rec                 535\n",
       "revol_bal               535\n",
       "revol_util              535\n",
       "total_acc               535\n",
       "mort_acc                535\n",
       "pub_rec_bankruptcies    535\n",
       "A                       535\n",
       "B                       535\n",
       "C                       535\n",
       "D                       535\n",
       "E                       535\n",
       "F                       535\n",
       "G                       535\n",
       "A1                      535\n",
       "A2                      535\n",
       "A3                      535\n",
       "A4                      535\n",
       "A5                      535\n",
       "B1                      535\n",
       "B2                      535\n",
       "B3                      535\n",
       "                       ... \n",
       "E5                      535\n",
       "F1                      535\n",
       "F2                      535\n",
       "F3                      535\n",
       "F4                      535\n",
       "F5                      535\n",
       "G1                      535\n",
       "G2                      535\n",
       "G3                      535\n",
       "G4                      535\n",
       "G5                      535\n",
       "Not Verified            535\n",
       "Source Verified         535\n",
       "Verified                535\n",
       "ANY                     535\n",
       "MORTGAGE                535\n",
       "NONE                    535\n",
       "OTHER                   535\n",
       "OWN                     535\n",
       "RENT                    535\n",
       "DIRECT_PAY              535\n",
       "INDIVIDUAL              535\n",
       "JOINT                   535\n",
       "acc_length              535\n",
       "ANY                       0\n",
       "MORTGAGE                  0\n",
       "NONE                      0\n",
       "OTHER                     0\n",
       "OWN                       0\n",
       "RENT                      0\n",
       "Length: 76, dtype: int64"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "# trying to fix nans in revol_util\n",
    "df.revol_util = df.revol_util.apply(lambda x: x if isinstance(x,float) else 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate the Neural Network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Test/Train Split the data\n",
    "\n",
    "need to use .values to turn the data into a numpy array from pandas values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "X = df.drop('loan_status',axis=1).values\n",
    "y = df['loan_status'].values\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=101)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### scale the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "# scale the data\n",
    "scaler = MinMaxScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = scaler.fit_transform(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0., 0., 0., ..., 0., 0., 0.])"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1., 0., 0., ..., 1., 0., 0.])"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### create the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(296621, 75)"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# match the columns row to the number of layers\n",
    "X_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### instantiate the model and add neural network layers\n",
    "\n",
    "In this cast we have 3 main layers, the first two neuron counts matching the columns, then third halving them and the last is an output neuron (1) as a boolean category (sigmoid function).\n",
    "\n",
    "In between each we do a dropout layer of 50%."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(75,activation='relu'))\n",
    "# configure the rate of drop: using 0 means no drop, 1 is all drops. Commonly .2 and .5.\n",
    "model.add(Dropout(0.5)) # %50 probability of being dropped\n",
    "model.add(Dense(75,activation='relu'))\n",
    "model.add(Dropout(0.5)) # %50 probability of being dropped\n",
    "model.add(Dense(31,activation='relu'))\n",
    "model.add(Dropout(0.5)) # %50 probability of being dropped\n",
    "model.add(Dense(1,activation='sigmoid')) # output neuron\n",
    "# since this is binary the loss function is binary_crossentropy\n",
    "model.compile(loss='binary_crossentropy',optimizer='adam')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create an early stop callback to cut down on overfitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "early_stop = EarlyStopping(monitor='val_loss',mode='min',patience=25,verbose=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fit the model using the early stop callback getting validation data\n",
    "\n",
    "This is now failing with the error:  \n",
    "\"ValueError: Failed to convert a NumPy array to a Tensor (Unsupported object type bool).\"\n",
    "\n",
    "Presuming this is due to converting loan_status to a boolean, so everything in Y is a boolean.\n",
    "array([False, False, False, ..., False, False, False], dtype=object)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4581 - val_loss: 0.4592\n",
      "Epoch 2/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4578 - val_loss: 0.4581\n",
      "Epoch 3/200\n",
      "9270/9270 [==============================] - 13s 1ms/step - loss: 0.4583 - val_loss: 0.4574\n",
      "Epoch 4/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4579 - val_loss: 0.4590\n",
      "Epoch 5/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4581 - val_loss: 0.4574\n",
      "Epoch 6/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4584 - val_loss: 0.4597\n",
      "Epoch 7/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4578 - val_loss: 0.4573\n",
      "Epoch 8/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4581 - val_loss: 0.4602\n",
      "Epoch 9/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4579 - val_loss: 0.4598\n",
      "Epoch 10/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4577 - val_loss: 0.4581\n",
      "Epoch 11/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4579 - val_loss: 0.4597\n",
      "Epoch 12/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4577 - val_loss: 0.4563\n",
      "Epoch 13/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4580 - val_loss: 0.4589\n",
      "Epoch 14/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4585 - val_loss: 0.4576\n",
      "Epoch 15/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4583 - val_loss: 0.4580\n",
      "Epoch 16/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4580 - val_loss: 0.4580\n",
      "Epoch 17/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4577 - val_loss: 0.4626\n",
      "Epoch 18/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4581 - val_loss: 0.4575\n",
      "Epoch 19/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4577 - val_loss: 0.4638\n",
      "Epoch 20/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4580 - val_loss: 0.4614\n",
      "Epoch 21/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4577 - val_loss: 0.4578\n",
      "Epoch 22/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4583 - val_loss: 0.4595\n",
      "Epoch 23/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4580 - val_loss: 0.4595\n",
      "Epoch 24/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4583 - val_loss: 0.4607\n",
      "Epoch 25/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4584 - val_loss: 0.4631\n",
      "Epoch 26/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4577 - val_loss: 0.4591\n",
      "Epoch 27/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4581 - val_loss: 0.4584\n",
      "Epoch 28/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4574 - val_loss: 0.4603\n",
      "Epoch 29/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4581 - val_loss: 0.4575\n",
      "Epoch 30/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4580 - val_loss: 0.4573\n",
      "Epoch 31/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4579 - val_loss: 0.4574\n",
      "Epoch 32/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4579 - val_loss: 0.4590\n",
      "Epoch 33/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4575 - val_loss: 0.4599\n",
      "Epoch 34/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4577 - val_loss: 0.4617\n",
      "Epoch 35/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4577 - val_loss: 0.4603\n",
      "Epoch 36/200\n",
      "9270/9270 [==============================] - 11s 1ms/step - loss: 0.4579 - val_loss: 0.4582\n",
      "Epoch 37/200\n",
      "9270/9270 [==============================] - 12s 1ms/step - loss: 0.4574 - val_loss: 0.4572\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History object at 0x16e0b8550>"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(x=X_train,y=y_train,epochs=200,validation_data=(X_test,y_test),verbose=1,\n",
    "         callbacks=[early_stop])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Analyze the loss in the model and find out where the early_stop killed the run. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot object at 0x16e05c090>"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lossdf = pd.DataFrame(model.history.history)\n",
    "lossdf.plot(figsize=(10,5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### predictions\n",
    "\n",
    "Gather predictions from the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0830 14:21:16.284619 4720309696 deprecation.py:323] From <ipython-input-121-bc83193b8b59>:1: Sequential.predict_classes (from tensorflow.python.keras.engine.sequential) is deprecated and will be removed after 2021-01-01.\n",
      "Instructions for updating:\n",
      "Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n"
     ]
    }
   ],
   "source": [
    "predictions = model.predict_classes(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Print the Classification Report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.80      1.00      0.89     79373\n",
      "         1.0       0.54      0.01      0.01     19501\n",
      "\n",
      "   micro avg       0.80      0.80      0.80     98874\n",
      "   macro avg       0.67      0.50      0.45     98874\n",
      "weighted avg       0.75      0.80      0.72     98874\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test,predictions))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Print a confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix\n",
      "[[79249   124]\n",
      " [19355   146]]\n",
      "True Negatives 79249 False Positives 124 False Negatives 19355 True Positives 146\n"
     ]
    }
   ],
   "source": [
    "# shows that the network only misclassified one point in the test data\n",
    "# True Negatives, False Positives\n",
    "# False Negatives, True Positives\n",
    "tn, fp, fn, tp = confusion_matrix(y_test,predictions).ravel()\n",
    "print(\"Confusion Matrix\")\n",
    "print(confusion_matrix(y_test,predictions))\n",
    "print('True Negatives',tn,'False Positives',fp,'False Negatives',fn,'True Positives',tp)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### test against known data\n",
    "\n",
    "grab the first index and drop the loan_status to make an X row"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "borrower = df.drop('loan_status',axis=1).iloc[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "75 values\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([10000.0, 36.0, 11.44, 329.48, 10, 117000.0, 26.24, 16.0, 0.0,\n",
       "       36369.0, 41.8, 25.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0,\n",
       "       0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0,\n",
       "       0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,\n",
       "       0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0,\n",
       "       0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.0, 2495.0, 0, 0, 0, 0, 0, 1],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(len(borrower), \"values\")\n",
    "borrower.values\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[10000.0, 36.0, 11.44, 329.48, 10, 117000.0, 26.24, 16.0, 0.0,\n",
       "        36369.0, 41.8, 25.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0,\n",
       "        0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0,\n",
       "        0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,\n",
       "        0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0,\n",
       "        0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.0, 2495.0, 0, 0, 0, 0, 0, 1]],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# scale the values\n",
    "borrower.values.reshape(-1,75)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "borrower = scaler.transform(borrower.values.reshape(-1,75))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.predict(borrower)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "borrower_predict = model.predict(borrower)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [],
   "source": [
    "borrower_actual = df['loan_status'].iloc[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### what does this mean?\n",
    "\n",
    " In this case I'm using a metrics I made up for the home prices (continuous data) that doesn't work well with Classification Data. Need something new here to be able to test on new data and see if the predictions are making sense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "actual vs predicted 0.0 [[0.16195416]]\n"
     ]
    }
   ],
   "source": [
    "print('actual vs predicted',borrower_actual,  borrower_predict)\n",
    "#print('swing percentage',borrower_actual / borrower_predict)\n",
    "#print('difference', borrower_predict - borrower_actual)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['loan_status'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a synthetic borrower:\n",
    "borrower2 = df.drop('loan_status',axis=1).iloc[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "loan_amnt                10000\n",
       "term                        36\n",
       "int_rate                 11.44\n",
       "installment             329.48\n",
       "emp_length                  10\n",
       "annual_inc              117000\n",
       "dti                      26.24\n",
       "open_acc                    16\n",
       "pub_rec                      0\n",
       "revol_bal                36369\n",
       "revol_util                41.8\n",
       "total_acc                   25\n",
       "mort_acc                     0\n",
       "pub_rec_bankruptcies         0\n",
       "A                            0\n",
       "B                            1\n",
       "C                            0\n",
       "D                            0\n",
       "E                            0\n",
       "F                            0\n",
       "G                            0\n",
       "A1                           0\n",
       "A2                           0\n",
       "A3                           0\n",
       "A4                           0\n",
       "A5                           0\n",
       "B1                           0\n",
       "B2                           0\n",
       "B3                           0\n",
       "B4                           1\n",
       "                         ...  \n",
       "E5                           0\n",
       "F1                           0\n",
       "F2                           0\n",
       "F3                           0\n",
       "F4                           0\n",
       "F5                           0\n",
       "G1                           0\n",
       "G2                           0\n",
       "G3                           0\n",
       "G4                           0\n",
       "G5                           0\n",
       "Not Verified                 1\n",
       "Source Verified              0\n",
       "Verified                     0\n",
       "ANY                          0\n",
       "MORTGAGE                     0\n",
       "NONE                         0\n",
       "OTHER                        0\n",
       "OWN                          0\n",
       "RENT                         1\n",
       "DIRECT_PAY                   0\n",
       "INDIVIDUAL                   1\n",
       "JOINT                        0\n",
       "acc_length                2495\n",
       "ANY                          0\n",
       "MORTGAGE                     0\n",
       "NONE                         0\n",
       "OTHER                        0\n",
       "OWN                          0\n",
       "RENT                         1\n",
       "Name: 0, Length: 75, dtype: object"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "borrower2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot object at 0x14e2b27d0>"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,10))\n",
    "sns.heatmap(df.corr(),cmap='plasma',linewidths=.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}