solution.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# The problem statement is given in the word file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b> Importing the dataset </b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>customer_id</th>\n",
       "      <th>vintage</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>dependents</th>\n",
       "      <th>occupation</th>\n",
       "      <th>city</th>\n",
       "      <th>customer_nw_category</th>\n",
       "      <th>branch_code</th>\n",
       "      <th>days_since_last_transaction</th>\n",
       "      <th>...</th>\n",
       "      <th>previous_month_end_balance</th>\n",
       "      <th>average_monthly_balance_prevQ</th>\n",
       "      <th>average_monthly_balance_prevQ2</th>\n",
       "      <th>current_month_credit</th>\n",
       "      <th>previous_month_credit</th>\n",
       "      <th>current_month_debit</th>\n",
       "      <th>previous_month_debit</th>\n",
       "      <th>current_month_balance</th>\n",
       "      <th>previous_month_balance</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>3135</td>\n",
       "      <td>66</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>187.0</td>\n",
       "      <td>2</td>\n",
       "      <td>755</td>\n",
       "      <td>224.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1449.07</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>310</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>3214</td>\n",
       "      <td>60.0</td>\n",
       "      <td>...</td>\n",
       "      <td>8704.66</td>\n",
       "      <td>7799.26</td>\n",
       "      <td>12419.41</td>\n",
       "      <td>0.56</td>\n",
       "      <td>0.56</td>\n",
       "      <td>5486.27</td>\n",
       "      <td>100.56</td>\n",
       "      <td>6496.78</td>\n",
       "      <td>8787.61</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4</td>\n",
       "      <td>2356</td>\n",
       "      <td>31</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>salaried</td>\n",
       "      <td>146.0</td>\n",
       "      <td>2</td>\n",
       "      <td>41</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>5815.29</td>\n",
       "      <td>4910.17</td>\n",
       "      <td>2815.94</td>\n",
       "      <td>0.61</td>\n",
       "      <td>0.61</td>\n",
       "      <td>6046.73</td>\n",
       "      <td>259.23</td>\n",
       "      <td>5006.28</td>\n",
       "      <td>5070.14</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>478</td>\n",
       "      <td>90</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>1020.0</td>\n",
       "      <td>2</td>\n",
       "      <td>582</td>\n",
       "      <td>147.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2291.91</td>\n",
       "      <td>2084.54</td>\n",
       "      <td>1006.54</td>\n",
       "      <td>0.47</td>\n",
       "      <td>0.47</td>\n",
       "      <td>0.47</td>\n",
       "      <td>2143.33</td>\n",
       "      <td>2291.91</td>\n",
       "      <td>1669.79</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6</td>\n",
       "      <td>2531</td>\n",
       "      <td>42</td>\n",
       "      <td>Male</td>\n",
       "      <td>2.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>1494.0</td>\n",
       "      <td>3</td>\n",
       "      <td>388</td>\n",
       "      <td>58.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1401.72</td>\n",
       "      <td>1643.31</td>\n",
       "      <td>1871.12</td>\n",
       "      <td>0.33</td>\n",
       "      <td>714.61</td>\n",
       "      <td>588.62</td>\n",
       "      <td>1538.06</td>\n",
       "      <td>1157.15</td>\n",
       "      <td>1677.16</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  vintage  age gender  dependents     occupation    city  \\\n",
       "0            1     3135   66   Male         0.0  self_employed   187.0   \n",
       "1            2      310   35   Male         0.0  self_employed     NaN   \n",
       "2            4     2356   31   Male         0.0       salaried   146.0   \n",
       "3            5      478   90    NaN         NaN  self_employed  1020.0   \n",
       "4            6     2531   42   Male         2.0  self_employed  1494.0   \n",
       "\n",
       "   customer_nw_category  branch_code  days_since_last_transaction  ...  \\\n",
       "0                     2          755                        224.0  ...   \n",
       "1                     2         3214                         60.0  ...   \n",
       "2                     2           41                          NaN  ...   \n",
       "3                     2          582                        147.0  ...   \n",
       "4                     3          388                         58.0  ...   \n",
       "\n",
       "   previous_month_end_balance  average_monthly_balance_prevQ  \\\n",
       "0                     1458.71                        1458.71   \n",
       "1                     8704.66                        7799.26   \n",
       "2                     5815.29                        4910.17   \n",
       "3                     2291.91                        2084.54   \n",
       "4                     1401.72                        1643.31   \n",
       "\n",
       "   average_monthly_balance_prevQ2  current_month_credit  \\\n",
       "0                         1449.07                  0.20   \n",
       "1                        12419.41                  0.56   \n",
       "2                         2815.94                  0.61   \n",
       "3                         1006.54                  0.47   \n",
       "4                         1871.12                  0.33   \n",
       "\n",
       "   previous_month_credit  current_month_debit  previous_month_debit  \\\n",
       "0                   0.20                 0.20                  0.20   \n",
       "1                   0.56              5486.27                100.56   \n",
       "2                   0.61              6046.73                259.23   \n",
       "3                   0.47                 0.47               2143.33   \n",
       "4                 714.61               588.62               1538.06   \n",
       "\n",
       "   current_month_balance  previous_month_balance  churn  \n",
       "0                1458.71                 1458.71      0  \n",
       "1                6496.78                 8787.61      0  \n",
       "2                5006.28                 5070.14      0  \n",
       "3                2291.91                 1669.79      1  \n",
       "4                1157.15                 1677.16      1  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data=pd.read_csv('churn_prediction.csv')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exploring the data and cleaning it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(28382, 21)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>customer_id</th>\n",
       "      <th>vintage</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>dependents</th>\n",
       "      <th>occupation</th>\n",
       "      <th>city</th>\n",
       "      <th>customer_nw_category</th>\n",
       "      <th>branch_code</th>\n",
       "      <th>days_since_last_transaction</th>\n",
       "      <th>...</th>\n",
       "      <th>previous_month_end_balance</th>\n",
       "      <th>average_monthly_balance_prevQ</th>\n",
       "      <th>average_monthly_balance_prevQ2</th>\n",
       "      <th>current_month_credit</th>\n",
       "      <th>previous_month_credit</th>\n",
       "      <th>current_month_debit</th>\n",
       "      <th>previous_month_debit</th>\n",
       "      <th>current_month_balance</th>\n",
       "      <th>previous_month_balance</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>28382.000000</td>\n",
       "      <td>28382.000000</td>\n",
       "      <td>28382.000000</td>\n",
       "      <td>27857</td>\n",
       "      <td>25919.000000</td>\n",
       "      <td>28302</td>\n",
       "      <td>27579.000000</td>\n",
       "      <td>28382.000000</td>\n",
       "      <td>28382.000000</td>\n",
       "      <td>25159.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>2.838200e+04</td>\n",
       "      <td>28382.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Male</td>\n",
       "      <td>NaN</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>16548</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17476</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>15143.508667</td>\n",
       "      <td>2364.336446</td>\n",
       "      <td>48.208336</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.347236</td>\n",
       "      <td>NaN</td>\n",
       "      <td>796.109576</td>\n",
       "      <td>2.225530</td>\n",
       "      <td>925.975019</td>\n",
       "      <td>69.997814</td>\n",
       "      <td>...</td>\n",
       "      <td>7.495771e+03</td>\n",
       "      <td>7.496780e+03</td>\n",
       "      <td>7.124209e+03</td>\n",
       "      <td>3.433252e+03</td>\n",
       "      <td>3.261694e+03</td>\n",
       "      <td>3.658745e+03</td>\n",
       "      <td>3.339761e+03</td>\n",
       "      <td>7.451133e+03</td>\n",
       "      <td>7.495177e+03</td>\n",
       "      <td>0.185329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>8746.454456</td>\n",
       "      <td>1610.124506</td>\n",
       "      <td>17.807163</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.997661</td>\n",
       "      <td>NaN</td>\n",
       "      <td>432.872102</td>\n",
       "      <td>0.660443</td>\n",
       "      <td>937.799129</td>\n",
       "      <td>86.341098</td>\n",
       "      <td>...</td>\n",
       "      <td>4.252935e+04</td>\n",
       "      <td>4.172622e+04</td>\n",
       "      <td>4.457581e+04</td>\n",
       "      <td>7.707145e+04</td>\n",
       "      <td>2.968889e+04</td>\n",
       "      <td>5.198542e+04</td>\n",
       "      <td>2.430111e+04</td>\n",
       "      <td>4.203394e+04</td>\n",
       "      <td>4.243198e+04</td>\n",
       "      <td>0.388571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>180.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>-3.149570e+03</td>\n",
       "      <td>1.428690e+03</td>\n",
       "      <td>-1.650610e+04</td>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>1.000000e-02</td>\n",
       "      <td>-3.374180e+03</td>\n",
       "      <td>-5.171920e+03</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>7557.250000</td>\n",
       "      <td>1121.000000</td>\n",
       "      <td>36.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>409.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>176.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.906000e+03</td>\n",
       "      <td>2.180945e+03</td>\n",
       "      <td>1.832507e+03</td>\n",
       "      <td>3.100000e-01</td>\n",
       "      <td>3.300000e-01</td>\n",
       "      <td>4.100000e-01</td>\n",
       "      <td>4.100000e-01</td>\n",
       "      <td>1.996765e+03</td>\n",
       "      <td>2.074408e+03</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>15150.500000</td>\n",
       "      <td>2018.000000</td>\n",
       "      <td>46.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>834.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>572.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>3.379915e+03</td>\n",
       "      <td>3.542865e+03</td>\n",
       "      <td>3.359600e+03</td>\n",
       "      <td>6.100000e-01</td>\n",
       "      <td>6.300000e-01</td>\n",
       "      <td>9.193000e+01</td>\n",
       "      <td>1.099600e+02</td>\n",
       "      <td>3.447995e+03</td>\n",
       "      <td>3.465235e+03</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>22706.750000</td>\n",
       "      <td>3176.000000</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1096.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1440.000000</td>\n",
       "      <td>95.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>6.656535e+03</td>\n",
       "      <td>6.666887e+03</td>\n",
       "      <td>6.517960e+03</td>\n",
       "      <td>7.072725e+02</td>\n",
       "      <td>7.492350e+02</td>\n",
       "      <td>1.360435e+03</td>\n",
       "      <td>1.357553e+03</td>\n",
       "      <td>6.667958e+03</td>\n",
       "      <td>6.654693e+03</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>30301.000000</td>\n",
       "      <td>12899.000000</td>\n",
       "      <td>90.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>52.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1649.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>4782.000000</td>\n",
       "      <td>365.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>5.740439e+06</td>\n",
       "      <td>5.700290e+06</td>\n",
       "      <td>5.010170e+06</td>\n",
       "      <td>1.226985e+07</td>\n",
       "      <td>2.361808e+06</td>\n",
       "      <td>7.637857e+06</td>\n",
       "      <td>1.414168e+06</td>\n",
       "      <td>5.778185e+06</td>\n",
       "      <td>5.720144e+06</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>11 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         customer_id       vintage           age gender    dependents  \\\n",
       "count   28382.000000  28382.000000  28382.000000  27857  25919.000000   \n",
       "unique           NaN           NaN           NaN      2           NaN   \n",
       "top              NaN           NaN           NaN   Male           NaN   \n",
       "freq             NaN           NaN           NaN  16548           NaN   \n",
       "mean    15143.508667   2364.336446     48.208336    NaN      0.347236   \n",
       "std      8746.454456   1610.124506     17.807163    NaN      0.997661   \n",
       "min         1.000000    180.000000      1.000000    NaN      0.000000   \n",
       "25%      7557.250000   1121.000000     36.000000    NaN      0.000000   \n",
       "50%     15150.500000   2018.000000     46.000000    NaN      0.000000   \n",
       "75%     22706.750000   3176.000000     60.000000    NaN      0.000000   \n",
       "max     30301.000000  12899.000000     90.000000    NaN     52.000000   \n",
       "\n",
       "           occupation          city  customer_nw_category   branch_code  \\\n",
       "count           28302  27579.000000          28382.000000  28382.000000   \n",
       "unique              5           NaN                   NaN           NaN   \n",
       "top     self_employed           NaN                   NaN           NaN   \n",
       "freq            17476           NaN                   NaN           NaN   \n",
       "mean              NaN    796.109576              2.225530    925.975019   \n",
       "std               NaN    432.872102              0.660443    937.799129   \n",
       "min               NaN      0.000000              1.000000      1.000000   \n",
       "25%               NaN    409.000000              2.000000    176.000000   \n",
       "50%               NaN    834.000000              2.000000    572.000000   \n",
       "75%               NaN   1096.000000              3.000000   1440.000000   \n",
       "max               NaN   1649.000000              3.000000   4782.000000   \n",
       "\n",
       "        days_since_last_transaction  ...  previous_month_end_balance  \\\n",
       "count                  25159.000000  ...                2.838200e+04   \n",
       "unique                          NaN  ...                         NaN   \n",
       "top                             NaN  ...                         NaN   \n",
       "freq                            NaN  ...                         NaN   \n",
       "mean                      69.997814  ...                7.495771e+03   \n",
       "std                       86.341098  ...                4.252935e+04   \n",
       "min                        0.000000  ...               -3.149570e+03   \n",
       "25%                       11.000000  ...                1.906000e+03   \n",
       "50%                       30.000000  ...                3.379915e+03   \n",
       "75%                       95.000000  ...                6.656535e+03   \n",
       "max                      365.000000  ...                5.740439e+06   \n",
       "\n",
       "        average_monthly_balance_prevQ  average_monthly_balance_prevQ2  \\\n",
       "count                    2.838200e+04                    2.838200e+04   \n",
       "unique                            NaN                             NaN   \n",
       "top                               NaN                             NaN   \n",
       "freq                              NaN                             NaN   \n",
       "mean                     7.496780e+03                    7.124209e+03   \n",
       "std                      4.172622e+04                    4.457581e+04   \n",
       "min                      1.428690e+03                   -1.650610e+04   \n",
       "25%                      2.180945e+03                    1.832507e+03   \n",
       "50%                      3.542865e+03                    3.359600e+03   \n",
       "75%                      6.666887e+03                    6.517960e+03   \n",
       "max                      5.700290e+06                    5.010170e+06   \n",
       "\n",
       "        current_month_credit  previous_month_credit  current_month_debit  \\\n",
       "count           2.838200e+04           2.838200e+04         2.838200e+04   \n",
       "unique                   NaN                    NaN                  NaN   \n",
       "top                      NaN                    NaN                  NaN   \n",
       "freq                     NaN                    NaN                  NaN   \n",
       "mean            3.433252e+03           3.261694e+03         3.658745e+03   \n",
       "std             7.707145e+04           2.968889e+04         5.198542e+04   \n",
       "min             1.000000e-02           1.000000e-02         1.000000e-02   \n",
       "25%             3.100000e-01           3.300000e-01         4.100000e-01   \n",
       "50%             6.100000e-01           6.300000e-01         9.193000e+01   \n",
       "75%             7.072725e+02           7.492350e+02         1.360435e+03   \n",
       "max             1.226985e+07           2.361808e+06         7.637857e+06   \n",
       "\n",
       "        previous_month_debit  current_month_balance  previous_month_balance  \\\n",
       "count           2.838200e+04           2.838200e+04            2.838200e+04   \n",
       "unique                   NaN                    NaN                     NaN   \n",
       "top                      NaN                    NaN                     NaN   \n",
       "freq                     NaN                    NaN                     NaN   \n",
       "mean            3.339761e+03           7.451133e+03            7.495177e+03   \n",
       "std             2.430111e+04           4.203394e+04            4.243198e+04   \n",
       "min             1.000000e-02          -3.374180e+03           -5.171920e+03   \n",
       "25%             4.100000e-01           1.996765e+03            2.074408e+03   \n",
       "50%             1.099600e+02           3.447995e+03            3.465235e+03   \n",
       "75%             1.357553e+03           6.667958e+03            6.654693e+03   \n",
       "max             1.414168e+06           5.778185e+06            5.720144e+06   \n",
       "\n",
       "               churn  \n",
       "count   28382.000000  \n",
       "unique           NaN  \n",
       "top              NaN  \n",
       "freq             NaN  \n",
       "mean        0.185329  \n",
       "std         0.388571  \n",
       "min         0.000000  \n",
       "25%         0.000000  \n",
       "50%         0.000000  \n",
       "75%         0.000000  \n",
       "max         1.000000  \n",
       "\n",
       "[11 rows x 21 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.describe(include='all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_id                          0\n",
       "vintage                              0\n",
       "age                                  0\n",
       "gender                             525\n",
       "dependents                        2463\n",
       "occupation                          80\n",
       "city                               803\n",
       "customer_nw_category                 0\n",
       "branch_code                          0\n",
       "days_since_last_transaction       3223\n",
       "current_balance                      0\n",
       "previous_month_end_balance           0\n",
       "average_monthly_balance_prevQ        0\n",
       "average_monthly_balance_prevQ2       0\n",
       "current_month_credit                 0\n",
       "previous_month_credit                0\n",
       "current_month_debit                  0\n",
       "previous_month_debit                 0\n",
       "current_month_balance                0\n",
       "previous_month_balance               0\n",
       "churn                                0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "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>customer_id</th>\n",
       "      <th>vintage</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>dependents</th>\n",
       "      <th>occupation</th>\n",
       "      <th>city</th>\n",
       "      <th>customer_nw_category</th>\n",
       "      <th>branch_code</th>\n",
       "      <th>days_since_last_transaction</th>\n",
       "      <th>...</th>\n",
       "      <th>previous_month_end_balance</th>\n",
       "      <th>average_monthly_balance_prevQ</th>\n",
       "      <th>average_monthly_balance_prevQ2</th>\n",
       "      <th>current_month_credit</th>\n",
       "      <th>previous_month_credit</th>\n",
       "      <th>current_month_debit</th>\n",
       "      <th>previous_month_debit</th>\n",
       "      <th>current_month_balance</th>\n",
       "      <th>previous_month_balance</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>3135</td>\n",
       "      <td>66</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>187.0</td>\n",
       "      <td>2</td>\n",
       "      <td>755</td>\n",
       "      <td>224.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1449.07</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>310</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>3214</td>\n",
       "      <td>60.0</td>\n",
       "      <td>...</td>\n",
       "      <td>8704.66</td>\n",
       "      <td>7799.26</td>\n",
       "      <td>12419.41</td>\n",
       "      <td>0.56</td>\n",
       "      <td>0.56</td>\n",
       "      <td>5486.27</td>\n",
       "      <td>100.56</td>\n",
       "      <td>6496.78</td>\n",
       "      <td>8787.61</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4</td>\n",
       "      <td>2356</td>\n",
       "      <td>31</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>salaried</td>\n",
       "      <td>146.0</td>\n",
       "      <td>2</td>\n",
       "      <td>41</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>5815.29</td>\n",
       "      <td>4910.17</td>\n",
       "      <td>2815.94</td>\n",
       "      <td>0.61</td>\n",
       "      <td>0.61</td>\n",
       "      <td>6046.73</td>\n",
       "      <td>259.23</td>\n",
       "      <td>5006.28</td>\n",
       "      <td>5070.14</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>478</td>\n",
       "      <td>90</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>1020.0</td>\n",
       "      <td>2</td>\n",
       "      <td>582</td>\n",
       "      <td>147.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2291.91</td>\n",
       "      <td>2084.54</td>\n",
       "      <td>1006.54</td>\n",
       "      <td>0.47</td>\n",
       "      <td>0.47</td>\n",
       "      <td>0.47</td>\n",
       "      <td>2143.33</td>\n",
       "      <td>2291.91</td>\n",
       "      <td>1669.79</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6</td>\n",
       "      <td>2531</td>\n",
       "      <td>42</td>\n",
       "      <td>Male</td>\n",
       "      <td>2.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>1494.0</td>\n",
       "      <td>3</td>\n",
       "      <td>388</td>\n",
       "      <td>58.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1401.72</td>\n",
       "      <td>1643.31</td>\n",
       "      <td>1871.12</td>\n",
       "      <td>0.33</td>\n",
       "      <td>714.61</td>\n",
       "      <td>588.62</td>\n",
       "      <td>1538.06</td>\n",
       "      <td>1157.15</td>\n",
       "      <td>1677.16</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28377</th>\n",
       "      <td>30297</td>\n",
       "      <td>1845</td>\n",
       "      <td>10</td>\n",
       "      <td>Female</td>\n",
       "      <td>0.0</td>\n",
       "      <td>student</td>\n",
       "      <td>1020.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1207</td>\n",
       "      <td>70.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1076.43</td>\n",
       "      <td>2282.19</td>\n",
       "      <td>2787.70</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.30</td>\n",
       "      <td>1076.43</td>\n",
       "      <td>1076.43</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28378</th>\n",
       "      <td>30298</td>\n",
       "      <td>4919</td>\n",
       "      <td>34</td>\n",
       "      <td>Female</td>\n",
       "      <td>0.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>1046.0</td>\n",
       "      <td>2</td>\n",
       "      <td>223</td>\n",
       "      <td>14.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4069.21</td>\n",
       "      <td>3668.83</td>\n",
       "      <td>3865.55</td>\n",
       "      <td>1.71</td>\n",
       "      <td>2.29</td>\n",
       "      <td>901.00</td>\n",
       "      <td>1014.07</td>\n",
       "      <td>3738.54</td>\n",
       "      <td>3690.32</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28379</th>\n",
       "      <td>30299</td>\n",
       "      <td>297</td>\n",
       "      <td>47</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>salaried</td>\n",
       "      <td>1096.0</td>\n",
       "      <td>2</td>\n",
       "      <td>588</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>61017.55</td>\n",
       "      <td>53444.81</td>\n",
       "      <td>21925.81</td>\n",
       "      <td>4666.84</td>\n",
       "      <td>3883.06</td>\n",
       "      <td>168.23</td>\n",
       "      <td>71.80</td>\n",
       "      <td>61078.50</td>\n",
       "      <td>57564.24</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28380</th>\n",
       "      <td>30300</td>\n",
       "      <td>2585</td>\n",
       "      <td>50</td>\n",
       "      <td>Male</td>\n",
       "      <td>3.0</td>\n",
       "      <td>self_employed</td>\n",
       "      <td>1219.0</td>\n",
       "      <td>3</td>\n",
       "      <td>274</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>1625.55</td>\n",
       "      <td>1683.20</td>\n",
       "      <td>1857.42</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>1625.55</td>\n",
       "      <td>1625.55</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28381</th>\n",
       "      <td>30301</td>\n",
       "      <td>2349</td>\n",
       "      <td>18</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.0</td>\n",
       "      <td>student</td>\n",
       "      <td>1232.0</td>\n",
       "      <td>2</td>\n",
       "      <td>474</td>\n",
       "      <td>59.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2821.34</td>\n",
       "      <td>3213.44</td>\n",
       "      <td>4447.45</td>\n",
       "      <td>0.11</td>\n",
       "      <td>7.44</td>\n",
       "      <td>714.40</td>\n",
       "      <td>1094.09</td>\n",
       "      <td>2402.62</td>\n",
       "      <td>3260.58</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>28382 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       customer_id  vintage  age  gender  dependents     occupation    city  \\\n",
       "0                1     3135   66    Male         0.0  self_employed   187.0   \n",
       "1                2      310   35    Male         0.0  self_employed     NaN   \n",
       "2                4     2356   31    Male         0.0       salaried   146.0   \n",
       "3                5      478   90     NaN         NaN  self_employed  1020.0   \n",
       "4                6     2531   42    Male         2.0  self_employed  1494.0   \n",
       "...            ...      ...  ...     ...         ...            ...     ...   \n",
       "28377        30297     1845   10  Female         0.0        student  1020.0   \n",
       "28378        30298     4919   34  Female         0.0  self_employed  1046.0   \n",
       "28379        30299      297   47    Male         0.0       salaried  1096.0   \n",
       "28380        30300     2585   50    Male         3.0  self_employed  1219.0   \n",
       "28381        30301     2349   18    Male         0.0        student  1232.0   \n",
       "\n",
       "       customer_nw_category  branch_code  days_since_last_transaction  ...  \\\n",
       "0                         2          755                        224.0  ...   \n",
       "1                         2         3214                         60.0  ...   \n",
       "2                         2           41                          NaN  ...   \n",
       "3                         2          582                        147.0  ...   \n",
       "4                         3          388                         58.0  ...   \n",
       "...                     ...          ...                          ...  ...   \n",
       "28377                     2         1207                         70.0  ...   \n",
       "28378                     2          223                         14.0  ...   \n",
       "28379                     2          588                          0.0  ...   \n",
       "28380                     3          274                          NaN  ...   \n",
       "28381                     2          474                         59.0  ...   \n",
       "\n",
       "       previous_month_end_balance  average_monthly_balance_prevQ  \\\n",
       "0                         1458.71                        1458.71   \n",
       "1                         8704.66                        7799.26   \n",
       "2                         5815.29                        4910.17   \n",
       "3                         2291.91                        2084.54   \n",
       "4                         1401.72                        1643.31   \n",
       "...                           ...                            ...   \n",
       "28377                     1076.43                        2282.19   \n",
       "28378                     4069.21                        3668.83   \n",
       "28379                    61017.55                       53444.81   \n",
       "28380                     1625.55                        1683.20   \n",
       "28381                     2821.34                        3213.44   \n",
       "\n",
       "       average_monthly_balance_prevQ2  current_month_credit  \\\n",
       "0                             1449.07                  0.20   \n",
       "1                            12419.41                  0.56   \n",
       "2                             2815.94                  0.61   \n",
       "3                             1006.54                  0.47   \n",
       "4                             1871.12                  0.33   \n",
       "...                               ...                   ...   \n",
       "28377                         2787.70                  0.30   \n",
       "28378                         3865.55                  1.71   \n",
       "28379                        21925.81               4666.84   \n",
       "28380                         1857.42                  0.20   \n",
       "28381                         4447.45                  0.11   \n",
       "\n",
       "       previous_month_credit  current_month_debit  previous_month_debit  \\\n",
       "0                       0.20                 0.20                  0.20   \n",
       "1                       0.56              5486.27                100.56   \n",
       "2                       0.61              6046.73                259.23   \n",
       "3                       0.47                 0.47               2143.33   \n",
       "4                     714.61               588.62               1538.06   \n",
       "...                      ...                  ...                   ...   \n",
       "28377                   0.30                 0.30                  0.30   \n",
       "28378                   2.29               901.00               1014.07   \n",
       "28379                3883.06               168.23                 71.80   \n",
       "28380                   0.20                 0.20                  0.20   \n",
       "28381                   7.44               714.40               1094.09   \n",
       "\n",
       "       current_month_balance  previous_month_balance  churn  \n",
       "0                    1458.71                 1458.71      0  \n",
       "1                    6496.78                 8787.61      0  \n",
       "2                    5006.28                 5070.14      0  \n",
       "3                    2291.91                 1669.79      1  \n",
       "4                    1157.15                 1677.16      1  \n",
       "...                      ...                     ...    ...  \n",
       "28377                1076.43                 1076.43      0  \n",
       "28378                3738.54                 3690.32      0  \n",
       "28379               61078.50                57564.24      1  \n",
       "28380                1625.55                 1625.55      0  \n",
       "28381                2402.62                 3260.58      1  \n",
       "\n",
       "[28382 rows x 21 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.drop_duplicates()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "************gender*************\n",
      "Male      16548\n",
      "Female    11309\n",
      "Name: gender, dtype: int64\n",
      "************dependents*************\n",
      "0.0     21435\n",
      "2.0      2150\n",
      "1.0      1395\n",
      "3.0       701\n",
      "4.0       179\n",
      "5.0        41\n",
      "6.0         8\n",
      "7.0         3\n",
      "36.0        1\n",
      "52.0        1\n",
      "25.0        1\n",
      "9.0         1\n",
      "50.0        1\n",
      "32.0        1\n",
      "8.0         1\n",
      "Name: dependents, dtype: int64\n",
      "Median : 0.0\n",
      "************occupation*************\n",
      "self_employed    17476\n",
      "salaried          6704\n",
      "student           2058\n",
      "retired           2024\n",
      "company             40\n",
      "Name: occupation, dtype: int64\n",
      "************city*************\n",
      "1020.0    3479\n",
      "1096.0    2016\n",
      "409.0     1334\n",
      "146.0     1291\n",
      "834.0     1138\n",
      "          ... \n",
      "629.0        1\n",
      "527.0        1\n",
      "1212.0       1\n",
      "530.0        1\n",
      "70.0         1\n",
      "Name: city, Length: 1604, dtype: int64\n",
      "Median : 834.0\n",
      "************days_since_last_transaction*************\n",
      "0.0      1672\n",
      "3.0       831\n",
      "14.0      654\n",
      "4.0       632\n",
      "6.0       596\n",
      "         ... \n",
      "170.0       1\n",
      "218.0       1\n",
      "334.0       1\n",
      "288.0       1\n",
      "321.0       1\n",
      "Name: days_since_last_transaction, Length: 360, dtype: int64\n",
      "Median : 30.0\n"
     ]
    }
   ],
   "source": [
    "missing = ['gender','dependents','occupation','city','days_since_last_transaction']\n",
    "\n",
    "for i in missing:\n",
    "    print(\"************\"+i+\"*************\")\n",
    "    print(data[i].value_counts())\n",
    "    if data[i].dtype != 'O':\n",
    "        print('Median : '+str(data[i].median()))\n",
    "        data[i].fillna(data[i].median(),inplace=True)\n",
    "    else:\n",
    "        data[i].fillna(data[i].mode()[0],inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_id                       0\n",
       "vintage                           0\n",
       "age                               0\n",
       "gender                            0\n",
       "dependents                        0\n",
       "occupation                        0\n",
       "city                              0\n",
       "customer_nw_category              0\n",
       "branch_code                       0\n",
       "days_since_last_transaction       0\n",
       "current_balance                   0\n",
       "previous_month_end_balance        0\n",
       "average_monthly_balance_prevQ     0\n",
       "average_monthly_balance_prevQ2    0\n",
       "current_month_credit              0\n",
       "previous_month_credit             0\n",
       "current_month_debit               0\n",
       "previous_month_debit              0\n",
       "current_month_balance             0\n",
       "previous_month_balance            0\n",
       "churn                             0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_id                         int64\n",
       "vintage                             int64\n",
       "age                                 int64\n",
       "gender                             object\n",
       "dependents                        float64\n",
       "occupation                         object\n",
       "city                              float64\n",
       "customer_nw_category                int64\n",
       "branch_code                         int64\n",
       "days_since_last_transaction       float64\n",
       "current_balance                   float64\n",
       "previous_month_end_balance        float64\n",
       "average_monthly_balance_prevQ     float64\n",
       "average_monthly_balance_prevQ2    float64\n",
       "current_month_credit              float64\n",
       "previous_month_credit             float64\n",
       "current_month_debit               float64\n",
       "previous_month_debit              float64\n",
       "current_month_balance             float64\n",
       "previous_month_balance            float64\n",
       "churn                               int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_id                       28382\n",
       "vintage                            5473\n",
       "age                                  90\n",
       "gender                                2\n",
       "dependents                           15\n",
       "occupation                            5\n",
       "city                               1604\n",
       "customer_nw_category                  3\n",
       "branch_code                        3185\n",
       "days_since_last_transaction         360\n",
       "current_balance                   27903\n",
       "previous_month_end_balance        27922\n",
       "average_monthly_balance_prevQ     27801\n",
       "average_monthly_balance_prevQ2    27940\n",
       "current_month_credit              10411\n",
       "previous_month_credit             10711\n",
       "current_month_debit               13704\n",
       "previous_month_debit              14010\n",
       "current_month_balance             27944\n",
       "previous_month_balance            27913\n",
       "churn                                 2\n",
       "dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "data=data.astype({'customer_nw_category':'object', 'dependents':'object', 'city':'object'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_id                         int64\n",
       "vintage                             int64\n",
       "age                                 int64\n",
       "gender                             object\n",
       "dependents                         object\n",
       "occupation                         object\n",
       "city                               object\n",
       "customer_nw_category               object\n",
       "branch_code                         int64\n",
       "days_since_last_transaction       float64\n",
       "current_balance                   float64\n",
       "previous_month_end_balance        float64\n",
       "average_monthly_balance_prevQ     float64\n",
       "average_monthly_balance_prevQ2    float64\n",
       "current_month_credit              float64\n",
       "previous_month_credit             float64\n",
       "current_month_debit               float64\n",
       "previous_month_debit              float64\n",
       "current_month_balance             float64\n",
       "previous_month_balance            float64\n",
       "churn                               int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "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>customer_id</th>\n",
       "      <th>vintage</th>\n",
       "      <th>age</th>\n",
       "      <th>branch_code</th>\n",
       "      <th>days_since_last_transaction</th>\n",
       "      <th>current_balance</th>\n",
       "      <th>previous_month_end_balance</th>\n",
       "      <th>average_monthly_balance_prevQ</th>\n",
       "      <th>average_monthly_balance_prevQ2</th>\n",
       "      <th>current_month_credit</th>\n",
       "      <th>previous_month_credit</th>\n",
       "      <th>current_month_debit</th>\n",
       "      <th>previous_month_debit</th>\n",
       "      <th>current_month_balance</th>\n",
       "      <th>previous_month_balance</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.006093</td>\n",
       "      <td>-0.000442</td>\n",
       "      <td>-0.000286</td>\n",
       "      <td>0.009699</td>\n",
       "      <td>0.006589</td>\n",
       "      <td>0.005819</td>\n",
       "      <td>0.004485</td>\n",
       "      <td>-0.002532</td>\n",
       "      <td>0.002494</td>\n",
       "      <td>-0.006414</td>\n",
       "      <td>0.002603</td>\n",
       "      <td>-0.008760</td>\n",
       "      <td>0.005140</td>\n",
       "      <td>0.004553</td>\n",
       "      <td>-0.002723</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>vintage</th>\n",
       "      <td>-0.006093</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.227390</td>\n",
       "      <td>-0.380714</td>\n",
       "      <td>-0.081342</td>\n",
       "      <td>0.002986</td>\n",
       "      <td>0.004518</td>\n",
       "      <td>0.003433</td>\n",
       "      <td>0.002244</td>\n",
       "      <td>0.002848</td>\n",
       "      <td>0.028034</td>\n",
       "      <td>0.005406</td>\n",
       "      <td>0.023949</td>\n",
       "      <td>0.001638</td>\n",
       "      <td>0.003448</td>\n",
       "      <td>-0.064637</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age</th>\n",
       "      <td>-0.000442</td>\n",
       "      <td>0.227390</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.058990</td>\n",
       "      <td>-0.012578</td>\n",
       "      <td>0.054346</td>\n",
       "      <td>0.058342</td>\n",
       "      <td>0.061708</td>\n",
       "      <td>0.059607</td>\n",
       "      <td>0.023840</td>\n",
       "      <td>0.029961</td>\n",
       "      <td>0.027702</td>\n",
       "      <td>0.033296</td>\n",
       "      <td>0.057662</td>\n",
       "      <td>0.060297</td>\n",
       "      <td>-0.020012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>branch_code</th>\n",
       "      <td>-0.000286</td>\n",
       "      <td>-0.380714</td>\n",
       "      <td>-0.058990</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.041643</td>\n",
       "      <td>0.000181</td>\n",
       "      <td>0.000214</td>\n",
       "      <td>0.001955</td>\n",
       "      <td>0.001310</td>\n",
       "      <td>-0.013988</td>\n",
       "      <td>-0.023849</td>\n",
       "      <td>-0.016944</td>\n",
       "      <td>-0.017584</td>\n",
       "      <td>0.001031</td>\n",
       "      <td>0.002080</td>\n",
       "      <td>0.035469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>days_since_last_transaction</th>\n",
       "      <td>0.009699</td>\n",
       "      <td>-0.081342</td>\n",
       "      <td>-0.012578</td>\n",
       "      <td>0.041643</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.017561</td>\n",
       "      <td>-0.012221</td>\n",
       "      <td>-0.010567</td>\n",
       "      <td>-0.005229</td>\n",
       "      <td>-0.030271</td>\n",
       "      <td>-0.064680</td>\n",
       "      <td>-0.035210</td>\n",
       "      <td>-0.062418</td>\n",
       "      <td>-0.015440</td>\n",
       "      <td>-0.009708</td>\n",
       "      <td>-0.040030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>current_balance</th>\n",
       "      <td>0.006589</td>\n",
       "      <td>0.002986</td>\n",
       "      <td>0.054346</td>\n",
       "      <td>0.000181</td>\n",
       "      <td>-0.017561</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.947276</td>\n",
       "      <td>0.958307</td>\n",
       "      <td>0.714600</td>\n",
       "      <td>0.030371</td>\n",
       "      <td>0.061754</td>\n",
       "      <td>0.044412</td>\n",
       "      <td>0.081247</td>\n",
       "      <td>0.983412</td>\n",
       "      <td>0.942207</td>\n",
       "      <td>-0.024181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>previous_month_end_balance</th>\n",
       "      <td>0.005819</td>\n",
       "      <td>0.004518</td>\n",
       "      <td>0.058342</td>\n",
       "      <td>0.000214</td>\n",
       "      <td>-0.012221</td>\n",
       "      <td>0.947276</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.970530</td>\n",
       "      <td>0.722998</td>\n",
       "      <td>0.032493</td>\n",
       "      <td>0.114222</td>\n",
       "      <td>0.066329</td>\n",
       "      <td>0.109606</td>\n",
       "      <td>0.974714</td>\n",
       "      <td>0.969605</td>\n",
       "      <td>0.006886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>average_monthly_balance_prevQ</th>\n",
       "      <td>0.004485</td>\n",
       "      <td>0.003433</td>\n",
       "      <td>0.061708</td>\n",
       "      <td>0.001955</td>\n",
       "      <td>-0.010567</td>\n",
       "      <td>0.958307</td>\n",
       "      <td>0.970530</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.763495</td>\n",
       "      <td>0.033639</td>\n",
       "      <td>0.085699</td>\n",
       "      <td>0.060579</td>\n",
       "      <td>0.121272</td>\n",
       "      <td>0.976290</td>\n",
       "      <td>0.994038</td>\n",
       "      <td>0.011960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>average_monthly_balance_prevQ2</th>\n",
       "      <td>-0.002532</td>\n",
       "      <td>0.002244</td>\n",
       "      <td>0.059607</td>\n",
       "      <td>0.001310</td>\n",
       "      <td>-0.005229</td>\n",
       "      <td>0.714600</td>\n",
       "      <td>0.722998</td>\n",
       "      <td>0.763495</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.036271</td>\n",
       "      <td>0.062264</td>\n",
       "      <td>0.045239</td>\n",
       "      <td>0.102519</td>\n",
       "      <td>0.725826</td>\n",
       "      <td>0.736635</td>\n",
       "      <td>0.018376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>current_month_credit</th>\n",
       "      <td>0.002494</td>\n",
       "      <td>0.002848</td>\n",
       "      <td>0.023840</td>\n",
       "      <td>-0.013988</td>\n",
       "      <td>-0.030271</td>\n",
       "      <td>0.030371</td>\n",
       "      <td>0.032493</td>\n",
       "      <td>0.033639</td>\n",
       "      <td>0.036271</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.168561</td>\n",
       "      <td>0.937021</td>\n",
       "      <td>0.135729</td>\n",
       "      <td>0.034182</td>\n",
       "      <td>0.038254</td>\n",
       "      <td>0.020755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>previous_month_credit</th>\n",
       "      <td>-0.006414</td>\n",
       "      <td>0.028034</td>\n",
       "      <td>0.029961</td>\n",
       "      <td>-0.023849</td>\n",
       "      <td>-0.064680</td>\n",
       "      <td>0.061754</td>\n",
       "      <td>0.114222</td>\n",
       "      <td>0.085699</td>\n",
       "      <td>0.062264</td>\n",
       "      <td>0.168561</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.165092</td>\n",
       "      <td>0.733953</td>\n",
       "      <td>0.085320</td>\n",
       "      <td>0.108496</td>\n",
       "      <td>0.042179</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>current_month_debit</th>\n",
       "      <td>0.002603</td>\n",
       "      <td>0.005406</td>\n",
       "      <td>0.027702</td>\n",
       "      <td>-0.016944</td>\n",
       "      <td>-0.035210</td>\n",
       "      <td>0.044412</td>\n",
       "      <td>0.066329</td>\n",
       "      <td>0.060579</td>\n",
       "      <td>0.045239</td>\n",
       "      <td>0.937021</td>\n",
       "      <td>0.165092</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.191755</td>\n",
       "      <td>0.069720</td>\n",
       "      <td>0.063375</td>\n",
       "      <td>0.048041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>previous_month_debit</th>\n",
       "      <td>-0.008760</td>\n",
       "      <td>0.023949</td>\n",
       "      <td>0.033296</td>\n",
       "      <td>-0.017584</td>\n",
       "      <td>-0.062418</td>\n",
       "      <td>0.081247</td>\n",
       "      <td>0.109606</td>\n",
       "      <td>0.121272</td>\n",
       "      <td>0.102519</td>\n",
       "      <td>0.135729</td>\n",
       "      <td>0.733953</td>\n",
       "      <td>0.191755</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.102010</td>\n",
       "      <td>0.139723</td>\n",
       "      <td>0.073058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>current_month_balance</th>\n",
       "      <td>0.005140</td>\n",
       "      <td>0.001638</td>\n",
       "      <td>0.057662</td>\n",
       "      <td>0.001031</td>\n",
       "      <td>-0.015440</td>\n",
       "      <td>0.983412</td>\n",
       "      <td>0.974714</td>\n",
       "      <td>0.976290</td>\n",
       "      <td>0.725826</td>\n",
       "      <td>0.034182</td>\n",
       "      <td>0.085320</td>\n",
       "      <td>0.069720</td>\n",
       "      <td>0.102010</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.963276</td>\n",
       "      <td>-0.006391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>previous_month_balance</th>\n",
       "      <td>0.004553</td>\n",
       "      <td>0.003448</td>\n",
       "      <td>0.060297</td>\n",
       "      <td>0.002080</td>\n",
       "      <td>-0.009708</td>\n",
       "      <td>0.942207</td>\n",
       "      <td>0.969605</td>\n",
       "      <td>0.994038</td>\n",
       "      <td>0.736635</td>\n",
       "      <td>0.038254</td>\n",
       "      <td>0.108496</td>\n",
       "      <td>0.063375</td>\n",
       "      <td>0.139723</td>\n",
       "      <td>0.963276</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.014593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>churn</th>\n",
       "      <td>-0.002723</td>\n",
       "      <td>-0.064637</td>\n",
       "      <td>-0.020012</td>\n",
       "      <td>0.035469</td>\n",
       "      <td>-0.040030</td>\n",
       "      <td>-0.024181</td>\n",
       "      <td>0.006886</td>\n",
       "      <td>0.011960</td>\n",
       "      <td>0.018376</td>\n",
       "      <td>0.020755</td>\n",
       "      <td>0.042179</td>\n",
       "      <td>0.048041</td>\n",
       "      <td>0.073058</td>\n",
       "      <td>-0.006391</td>\n",
       "      <td>0.014593</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                customer_id   vintage       age  branch_code  \\\n",
       "customer_id                        1.000000 -0.006093 -0.000442    -0.000286   \n",
       "vintage                           -0.006093  1.000000  0.227390    -0.380714   \n",
       "age                               -0.000442  0.227390  1.000000    -0.058990   \n",
       "branch_code                       -0.000286 -0.380714 -0.058990     1.000000   \n",
       "days_since_last_transaction        0.009699 -0.081342 -0.012578     0.041643   \n",
       "current_balance                    0.006589  0.002986  0.054346     0.000181   \n",
       "previous_month_end_balance         0.005819  0.004518  0.058342     0.000214   \n",
       "average_monthly_balance_prevQ      0.004485  0.003433  0.061708     0.001955   \n",
       "average_monthly_balance_prevQ2    -0.002532  0.002244  0.059607     0.001310   \n",
       "current_month_credit               0.002494  0.002848  0.023840    -0.013988   \n",
       "previous_month_credit             -0.006414  0.028034  0.029961    -0.023849   \n",
       "current_month_debit                0.002603  0.005406  0.027702    -0.016944   \n",
       "previous_month_debit              -0.008760  0.023949  0.033296    -0.017584   \n",
       "current_month_balance              0.005140  0.001638  0.057662     0.001031   \n",
       "previous_month_balance             0.004553  0.003448  0.060297     0.002080   \n",
       "churn                             -0.002723 -0.064637 -0.020012     0.035469   \n",
       "\n",
       "                                days_since_last_transaction  current_balance  \\\n",
       "customer_id                                        0.009699         0.006589   \n",
       "vintage                                           -0.081342         0.002986   \n",
       "age                                               -0.012578         0.054346   \n",
       "branch_code                                        0.041643         0.000181   \n",
       "days_since_last_transaction                        1.000000        -0.017561   \n",
       "current_balance                                   -0.017561         1.000000   \n",
       "previous_month_end_balance                        -0.012221         0.947276   \n",
       "average_monthly_balance_prevQ                     -0.010567         0.958307   \n",
       "average_monthly_balance_prevQ2                    -0.005229         0.714600   \n",
       "current_month_credit                              -0.030271         0.030371   \n",
       "previous_month_credit                             -0.064680         0.061754   \n",
       "current_month_debit                               -0.035210         0.044412   \n",
       "previous_month_debit                              -0.062418         0.081247   \n",
       "current_month_balance                             -0.015440         0.983412   \n",
       "previous_month_balance                            -0.009708         0.942207   \n",
       "churn                                             -0.040030        -0.024181   \n",
       "\n",
       "                                previous_month_end_balance  \\\n",
       "customer_id                                       0.005819   \n",
       "vintage                                           0.004518   \n",
       "age                                               0.058342   \n",
       "branch_code                                       0.000214   \n",
       "days_since_last_transaction                      -0.012221   \n",
       "current_balance                                   0.947276   \n",
       "previous_month_end_balance                        1.000000   \n",
       "average_monthly_balance_prevQ                     0.970530   \n",
       "average_monthly_balance_prevQ2                    0.722998   \n",
       "current_month_credit                              0.032493   \n",
       "previous_month_credit                             0.114222   \n",
       "current_month_debit                               0.066329   \n",
       "previous_month_debit                              0.109606   \n",
       "current_month_balance                             0.974714   \n",
       "previous_month_balance                            0.969605   \n",
       "churn                                             0.006886   \n",
       "\n",
       "                                average_monthly_balance_prevQ  \\\n",
       "customer_id                                          0.004485   \n",
       "vintage                                              0.003433   \n",
       "age                                                  0.061708   \n",
       "branch_code                                          0.001955   \n",
       "days_since_last_transaction                         -0.010567   \n",
       "current_balance                                      0.958307   \n",
       "previous_month_end_balance                           0.970530   \n",
       "average_monthly_balance_prevQ                        1.000000   \n",
       "average_monthly_balance_prevQ2                       0.763495   \n",
       "current_month_credit                                 0.033639   \n",
       "previous_month_credit                                0.085699   \n",
       "current_month_debit                                  0.060579   \n",
       "previous_month_debit                                 0.121272   \n",
       "current_month_balance                                0.976290   \n",
       "previous_month_balance                               0.994038   \n",
       "churn                                                0.011960   \n",
       "\n",
       "                                average_monthly_balance_prevQ2  \\\n",
       "customer_id                                          -0.002532   \n",
       "vintage                                               0.002244   \n",
       "age                                                   0.059607   \n",
       "branch_code                                           0.001310   \n",
       "days_since_last_transaction                          -0.005229   \n",
       "current_balance                                       0.714600   \n",
       "previous_month_end_balance                            0.722998   \n",
       "average_monthly_balance_prevQ                         0.763495   \n",
       "average_monthly_balance_prevQ2                        1.000000   \n",
       "current_month_credit                                  0.036271   \n",
       "previous_month_credit                                 0.062264   \n",
       "current_month_debit                                   0.045239   \n",
       "previous_month_debit                                  0.102519   \n",
       "current_month_balance                                 0.725826   \n",
       "previous_month_balance                                0.736635   \n",
       "churn                                                 0.018376   \n",
       "\n",
       "                                current_month_credit  previous_month_credit  \\\n",
       "customer_id                                 0.002494              -0.006414   \n",
       "vintage                                     0.002848               0.028034   \n",
       "age                                         0.023840               0.029961   \n",
       "branch_code                                -0.013988              -0.023849   \n",
       "days_since_last_transaction                -0.030271              -0.064680   \n",
       "current_balance                             0.030371               0.061754   \n",
       "previous_month_end_balance                  0.032493               0.114222   \n",
       "average_monthly_balance_prevQ               0.033639               0.085699   \n",
       "average_monthly_balance_prevQ2              0.036271               0.062264   \n",
       "current_month_credit                        1.000000               0.168561   \n",
       "previous_month_credit                       0.168561               1.000000   \n",
       "current_month_debit                         0.937021               0.165092   \n",
       "previous_month_debit                        0.135729               0.733953   \n",
       "current_month_balance                       0.034182               0.085320   \n",
       "previous_month_balance                      0.038254               0.108496   \n",
       "churn                                       0.020755               0.042179   \n",
       "\n",
       "                                current_month_debit  previous_month_debit  \\\n",
       "customer_id                                0.002603             -0.008760   \n",
       "vintage                                    0.005406              0.023949   \n",
       "age                                        0.027702              0.033296   \n",
       "branch_code                               -0.016944             -0.017584   \n",
       "days_since_last_transaction               -0.035210             -0.062418   \n",
       "current_balance                            0.044412              0.081247   \n",
       "previous_month_end_balance                 0.066329              0.109606   \n",
       "average_monthly_balance_prevQ              0.060579              0.121272   \n",
       "average_monthly_balance_prevQ2             0.045239              0.102519   \n",
       "current_month_credit                       0.937021              0.135729   \n",
       "previous_month_credit                      0.165092              0.733953   \n",
       "current_month_debit                        1.000000              0.191755   \n",
       "previous_month_debit                       0.191755              1.000000   \n",
       "current_month_balance                      0.069720              0.102010   \n",
       "previous_month_balance                     0.063375              0.139723   \n",
       "churn                                      0.048041              0.073058   \n",
       "\n",
       "                                current_month_balance  previous_month_balance  \\\n",
       "customer_id                                  0.005140                0.004553   \n",
       "vintage                                      0.001638                0.003448   \n",
       "age                                          0.057662                0.060297   \n",
       "branch_code                                  0.001031                0.002080   \n",
       "days_since_last_transaction                 -0.015440               -0.009708   \n",
       "current_balance                              0.983412                0.942207   \n",
       "previous_month_end_balance                   0.974714                0.969605   \n",
       "average_monthly_balance_prevQ                0.976290                0.994038   \n",
       "average_monthly_balance_prevQ2               0.725826                0.736635   \n",
       "current_month_credit                         0.034182                0.038254   \n",
       "previous_month_credit                        0.085320                0.108496   \n",
       "current_month_debit                          0.069720                0.063375   \n",
       "previous_month_debit                         0.102010                0.139723   \n",
       "current_month_balance                        1.000000                0.963276   \n",
       "previous_month_balance                       0.963276                1.000000   \n",
       "churn                                       -0.006391                0.014593   \n",
       "\n",
       "                                   churn  \n",
       "customer_id                    -0.002723  \n",
       "vintage                        -0.064637  \n",
       "age                            -0.020012  \n",
       "branch_code                     0.035469  \n",
       "days_since_last_transaction    -0.040030  \n",
       "current_balance                -0.024181  \n",
       "previous_month_end_balance      0.006886  \n",
       "average_monthly_balance_prevQ   0.011960  \n",
       "average_monthly_balance_prevQ2  0.018376  \n",
       "current_month_credit            0.020755  \n",
       "previous_month_credit           0.042179  \n",
       "current_month_debit             0.048041  \n",
       "previous_month_debit            0.073058  \n",
       "current_month_balance          -0.006391  \n",
       "previous_month_balance          0.014593  \n",
       "churn                           1.000000  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# correlation\n",
    "data.corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b> Univariate Analysis using Box Plot </b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWoAAAEICAYAAAB25L6yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAASJElEQVR4nO3df3TVd33H8deLJCVrqC3YuKmUohvrUrKKLsce1+qkcEaLTp2nU3Omo4NTlnN2cvDYDZScs+F24GycDuWkzgyEOU4lVTudnh6psEN6MFp/BAs9tKlDXVlzqiPaohJGStr3/rhfMMQL+Qbyzf0keT7Oucd7v9/v/d5XburrfPjcT77XESEAQLpmVDoAAODiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1ACQOIoa047tDbbvv8Tn3mW7e7wzARdDUWNKsP1W232VzgEUgaLGhHPJjBHbqiuVB0gdRY0xs32d7S/Y7rf9U9v3jZxOsD3fdpwtYNuP2N5o++uSTkl6bbb/L20flXQ0O+7ttg/ZPmH7G7ZvGnbOp23/le3Hbf/M9mdt19quk7RH0qtsn8xurxrlx6jNnv8L29+1/bphr/Nh2z/I9j1p+48v8l5stf2M7Z/bPmj7zcP2bbD9Odu7snM9YbvpYu/jsH0rbffaft72V21fP/pvBlNVYUVte6ft47aP5Dz+Pdn/KZ6wvbuoXLg8tqskPSTpmKT5kl4t6YGcT/+ApNWSrsqeL0nvknSzpBttv0HSTkl/Ienlkv5F0pdtzxx2jvdIul3SayTdJOmuiBiQdIekZyNiVnZ7dpQs75T0eUlzJO2W9B+2a7J9P5D0ZklXS/qopPttv/IC5/mOpEXDzvN527XD9r9DpffnGklflnSfdPH30fa7JK2X9G5J9ZK+JqlzlJ8HU1lEFHKT9BZJb5B0JMexCyQ9Jml29vgVReXidtm/1zdJ6pdUPWL7Bkn3D3s8X1KcPU7SI5L+bsRzQtJtwx5/UtLfjzjme5L+ILv/tKT3D9u3WVJHdv+tkvpy/gwbJH1z2OMZkn4k6c0XOP6QpHdm9++S1H2Rcz8v6XXDXuc/h+27UdL/Xex9zPbtkbRqRL5Tkq6v9O+fW2VuhY2oI+KApOeGb7P9m7Yfzv6J+DXbv5PtulvSJyLi+ey5x4vKhct2naRjETF0Cc99ZpRt10u6J5v2OGH7RPZ6w6cxfjzs/ilJsy4hx3mvGxEvSeo7+zq2/2zY9MsJSY2Sri13Etv3ZFMUP8uOvXrEsSPz1mbTQRd7H6+XtHXY6z8nySqNujENTfQHONsktUTEUds3S/pnSbdJ+m1JyuYvqyRtiIiHJzgb8nlG0jzb1SNKZkDSlcMe/0aZ55a7pu7wbc9I2hgRGy8h11iv13vd2TvZB5tzJT2bzQVvl7RE0qMR8aLtQyoV5Xmy+eh12bFPRMRLtp8vd2wZF3ofz+7bGBGfGePPhClqwj5MtD1L0u+rNId3SKX5x7PzftUqTX+8VVKzpE/ZvmaismFMvq3SNME/2K7LPsy7RaXpgbfYnmf7akkfuYRzb5fUYvvmbGVIne232b4qx3P/V9LLs9fO4/dsvzsb3X5Q0qCkb0qqU6n0+yXJ9p+rNKIu5ypJQ9mx1bb/RtLLcr7+hd5HSeqQ9BHbC7MMV9v+k5znxRQ0kas+Zkg6ERGLht0asn19kr4UEWci4r9VmpdcMIHZkFNEvCjpjyT9lqT/Uel3996I2Cfps5Iel3RQpQ/KxnruHpWmwe5Taa73+yrNCed57lMqfeD2w2zKYLRVH1+S9N7sdT4g6d3Zf39PSvonSY+qVP6/K+nrFzjHV1WaT/4vlT4UPK3y0zvl8pZ9H7N9X5T0j5IesP1zSUdU+rAU05QjivuGF9vzJT0UEY3Z429I+lhEfN62Jd0UEYdt3y6pOSJW2L5WpQ8WF0XETwsLBwCTRJHL8zpVGpXcYLvP9ipJfypple3Dkp5QaYmUVBqZ/NT2k5K6JP01JQ0AJYWOqIFKsb1HpbXQI22KiE0TnQe4HBQ1ACSukOV51157bcyfP7+IUwPAlHTw4MGfRER9uX2FFPX8+fPV09NTxKkBYEqyfexC+7goEwAkjqIGgMTlKmrb19h+0PZT2XUN3lR0MABASd456q2SHo6IO21fofOv6QAAKNCoRW37ZSpdsvQuSYqIFyS9UGwsAMBZeaY+XqvSRWf+1fZjtj/l0jdqnMf2ats9tnv6+/vHPShwOTo7O9XY2Kiqqio1Njaqs5Pr8GPyyFPU1Sp9AcAnI+L1Kl3O8sMjD4qIbRHRFBFN9fVllwICFdHZ2am2tja1t7fr9OnTam9vV1tbG2WNSSNPUfep9M0Z38oeP6hScQOTwsaNG7Vjxw4tXrxYNTU1Wrx4sXbs2KGNGy/lstfAxBu1qCPix5KesX1DtmmJpCcLTQWMo97eXt16663nbbv11lvV29tboUTA2ORdR90q6TO2H1fpizy5qA0mjYaGBnV3d5+3rbu7Ww0NDRd4BpCWXEUdEYey+eebIuJdZ7/bEJgM2tratGrVKnV1denMmTPq6urSqlWr1NbWVuloQC4T/Z2JwIRrbm6WJLW2tqq3t1cNDQ3auHHjue1A6gq5zGlTU1NwUSYAyM/2wYhoKrePa30AQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEgcRQ0AiaOoASBxFDUAJI6iBoDEVec5yPbTkn4h6UVJQxf6SnMAwPjLVdSZxRHxk8KSAADKYuoDABKXt6hD0l7bB22vLjIQAOB8eac+bomIZ22/QtI+209FxIHhB2QFvlqS5s2bN84xAWD6yjWijohns/89LumLkt5Y5phtEdEUEU319fXjmxIAprFRi9p2ne2rzt6X9IeSjhQdDABQkmfq49clfdH22eN3R8TDhaYCAJwzalFHxA8lvW4CsgAAymB5HgAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEgcRQ0AiaOoASBxFDUAJI6iBoDEUdQAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkLndR266y/Zjth4oMBAA431hG1Gsk9RYVBABQXq6itj1X0tskfarYOACAkfKOqD8uaa2kly50gO3Vtnts9/T3949LOABAjqK2/XZJxyPi4MWOi4htEdEUEU319fXjFhAAprs8I+pbJL3D9tOSHpB0m+37C00FADhn1KKOiI9ExNyImC/pfZL2R8T7C08GAJDEOmoASF71WA6OiEckPVJIEgBAWYyoASBxFDUAJI6iBoDEUdQAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEgcRQ0AiRu1qG3X2v627cO2n7D90YkIBgAoqc5xzKCk2yLipO0aSd2290TENwvOBgBQjqKOiJB0MntYk92iyFAAgF/KNUdtu8r2IUnHJe2LiG+VOWa17R7bPf39/eOdEwCmrVxFHREvRsQiSXMlvdF2Y5ljtkVEU0Q01dfXj3dOAJi2xrTqIyJOSHpE0u2FpAEA/Io8qz7qbV+T3f81SUslPVV0MABASZ5VH6+U9G+2q1Qq9s9FxEPFxgIAnJVn1cfjkl4/AVkAAGXwl4kAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEgcRQ0AiaOoASBxFDUAJI6iBoDEUdQAkLhRi9r2dba7bPfafsL2mokIBgAoqc5xzJCkeyLiu7avknTQ9r6IeLLgbAAA5RhRR8SPIuK72f1fSOqV9OqigwEASsY0R217vqTXS/pWmX2rbffY7unv7x+fdACA/EVte5akf5f0wYj4+cj9EbEtIpoioqm+vn48MwLAtJarqG3XqFTSn4mILxQbCQAwXJ5VH5a0Q1JvRGwpPhIAYLg8I+pbJH1A0m22D2W35QXnAgBkRl2eFxHdkjwBWQAAZfCXiQCQOIoaABJHUQNA4ihqAEgcRQ0AiaOoASBxFDUAJI6iBoDEUdQAkDiKGgASR1FjWujs7FRjY6OqqqrU2Niozs7OSkcCcqOoMeV1dnZqzZo1GhgYUERoYGBAa9asoawxaVDUmPLWrl2rqqoq7dy5U4ODg9q5c6eqqqq0du3aSkcDcqGoMeX19fVp165dWrx4sWpqarR48WLt2rVLfX19lY4G5EJRA0DiKGpMeXPnztWKFSvU1dWlM2fOqKurSytWrNDcuXMrHQ3IhaLGlLd582YNDQ1p5cqVqq2t1cqVKzU0NKTNmzdXOhqQC0WNKa+5uVlbt25VXV2dJKmurk5bt25Vc3NzhZMB+Tgixv2kTU1N0dPTM+7nBYCpyvbBiGgqt48RNQAkjqIGgMRR1ACQOIoa00Jra6tqa2tlW7W1tWptba10JCA3ihpTXmtrqzo6OrRp0yYNDAxo06ZN6ujooKwxaYy66sP2Tklvl3Q8IhrznJRVH0hJbW2t7rzzTh06dEi9vb1qaGjQokWL9OCDD+r06dOVjgdIuvxVH5+WdPu4JgIm0ODgoLq7u9Xe3q7Tp0+rvb1d3d3dGhwcrHQ0IJdRizoiDkh6bgKyAIWwreXLl593Uably5fLdqWjAbmM2xy17dW2e2z39Pf3j9dpgcsWEdq+fbu2bNmiU6dOacuWLdq+fbuK+GMvoAi5/jLR9nxJDzFHjcmosbFRCxYs0J49ezQ4OKiZM2fqjjvu0NGjR3XkyJFKxwMk8ZeJmOba2tp0+PBh7dmzRy+88IL27Nmjw4cPq62trdLRgFwoakx5zc3NWrBggZYsWaIrrrhCS5Ys0YIFC7goEyaNUYvadqekRyXdYLvP9qriYwHjp7W1Vfv379e9996rgYEB3Xvvvdq/fz/rqDFpcPU8THm1tbXatGmTPvShD53btmXLFq1fv5511EjGxeaoKWpMebY1MDCgK6+88ty2U6dOqa6ujpUfSAYfJmJamzlzpjo6Os7b1tHRoZkzZ1YoETA21ZUOABTt7rvv1rp16yRJLS0t6ujo0Lp169TS0lLhZEA+FDWmvPb2dknS+vXrdc8992jmzJlqaWk5tx1IHXPUAJAA5qgx7S1btkwzZsyQbc2YMUPLli2rdCQgN4oaU96yZcu0d+9etbS06MSJE2ppadHevXspa0wazFFjytu3b5+WLl2qAwcOaM6cOWpoaNDSpUu1b9++SkcDcmFEjSkvInT06NHzrkd99OhR1lBj0mBEjWlhzpw5am1tPfcNL3PmzNGxY8cqHQvIhaLGtPDYY49p1qxZeumll3Ts2DGdPHmy0pGA3Jj6wJRXXV0t2+fK+eTJk7Kt6mrGKZgcKGpMeUNDQ4oIzZ49W5I0e/ZsRYSGhoYqnAzIh6LGtFBTU3PeiLqmpqbCiYD8+LcfpoUzZ86UvQ9MBoyoASBxFDUAJI6iBoDEUdQAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcRQ1ACQuV1Hbvt3292x/3/aHiw4FAPilUYvadpWkT0i6Q9KNkppt31h0MABASZ6LMr1R0vcj4oeSZPsBSe+U9GSRwYDR2J6wc/C1XaikPEX9aknPDHvcJ+nmkQfZXi1ptSTNmzdvXMJhmtlw9ZgOj799WUFByhhjNm34WTE5MC3lKepyQ45fGV5ExDZJ2ySpqamJ4QfGrqByu9iomZEyJoM8Hyb2Sbpu2OO5kp4tJg4w/nbv3j2m7UBq8hT1dyQtsP0a21dIep+kLxcbCxg/zc3N2r17txYuXKgZM2Zo4cKF2r17t5qbmysdDcjFef7pZ3u5pI9LqpK0MyI2Xuz4pqam6OnpGZ+EADAN2D4YEU3l9uX6Kq6I+Iqkr4xrKgBALvxlIgAkjqIGgMRR1ACQOIoaABKXa9XHmE9q90s6Nu4nBi7ftZJ+UukQQBnXR0R9uR2FFDWQKts9F1oCBaSKqQ8ASBxFDQCJo6gx3WyrdABgrJijBoDEMaIGgMRR1ACQOIoa04LtnbaP2z5S6SzAWFHUmC4+Len2SocALgVFjWkhIg5Ieq7SOYBLQVEDQOIoagBIHEUNAImjqAEgcRQ1pgXbnZIelXSD7T7bqyqdCciLPyEHgMQxogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHH/D3z+fJvoSEpvAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWoAAAEICAYAAAB25L6yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAVtElEQVR4nO3df5TddX3n8eebSUwgREJg1EhI0oI/hk4r3Z1Cj7K28UcPUM9K15aaVgWZQuXUNN3+oLKzZ9UtOXuWs1lFZBepSWNqmYLb6lJKrSkOpSMqBqU0P3RFGkqUH/nJj+ikSXjvH9/vhEuYZG6SubmfufN8nDNnvvf7832/985rPt/P93u/NzITSVK5Tmh3AZKkwzOoJalwBrUkFc6glqTCGdSSVDiDWpIKZ1Br0oqIjIizDzHt8ogYPsb1b46Itx3lsvdExG8cy/alUQa1JgWDb3KKiHMi4o6IeDoino2IL0fEz7a7rsnGoO4AETGt3TWo80TlqDMiIs4CvgL8E/BjwKuBLwBrI+K8ialyajCoj4OI+FBEfK9uUWyMiF+KiBkRsSsiehvm646IH0XEK+rH74iIB+v57ouIn2qYd3NE/GFEPATsjohpY22nYf6uiFgREdsi4p8j4oN118G0evopEbEyIh6PiO9HxHUR0TXO87o8Ir4SER+ra3wkIt5Yj38sIp6KiMsa5j8lItZExNaIeDQi/vNoEIx2VUTE/4iInXWNF9XTlgP/DvhkRDwXEZ9sKONtEfHdepmbIiLGqPOmiFhx0Li/iojfaeLl+5l6X+6MiD+JiJn18qdGxJ31c9lZD88/xH46q25Jbq/3/59FxJyG6Zsj4vcj4qG65Xnb6Hbq6e+s3wfP1K/vhQ3782hfsxvrbX07It7aMP2eiFgeEV8Bfgj8eES8PiLWRsSOiPhORFxaz/uzEfFE4zbr9/ZD9cOPAF/NzIHM3JGZz2bmJ4DPAv+9iX2vUZnpT4t/gF+hak2cAPwqsBuYB6wCljfM91vAF+vhfwM8BZwPdAGXAZuBGfX0zcCDwJnAiYfbTj3tA8BGYD5wKvB3QALT6ulfAD4FzAJeAdwP/OY4z+tyYB/w/rrG64B/AW4CZgC/ADwLnFzPvwb4v8BsYBHw/4D+hnXtBa6s13U18AMg6un3AL9x0PYTuBOYAywAtgIXNqxvuB4+r17XCfXj06lC6JXjPL/NwPp6H8+lah1eV087DXgXcFL9fD4HfKFh2QP1AmcDb6/3STdwL/Dxg7Zzf/3azQU2AR9oqP3pevkTgDOA10/Aa/Yfgen1++RpYG5D3f8C/AQwDTgFeKx+jadRvS+3AT9Rz/894O0N6/8c8KF6+Ang/WPUsLiuYWa7/zYny08rw2kVVdCsb3L+S6mCZANwa7t3TEt3ehWw7wTeBjzSMP4rwPvq4f8N/NFBy30H+Ll6eDNwRTPbqYe/3PhHXG876z++VwJ7qAO/nr4EGBpn/ZcD3214/JP1Ol/ZMG47cC5V+O4BzmmY9pvAPQ3rerhh2kn1ul5VP76HsYP6gobHtzeExOXUQV0/3jQaKMAHgbuaeJ02Uwdm/fhi4HuHmPdcYGfD45fU2zDtEuBbB23nPQ2Prwduroc/BXxsjHUcy2t24B9gPe5+4L0Ndf/Xhmm/CvzDQev4FPDhevg6YFU9PJuqcbCwfryP+h/nQcu/vn7tXt3Kv7NO+mll3+Zq4JNUrajDiojXANcCb8rMnaOH/p0iIt4H/C5VKxLgZKpW3V8BJ0bE+VStj3OBz9fzLAQui4ilDat6GVWra9RjTW6HernG+RuHF1K1rh5v6Dk44eD1H8KTDcM/AsjMg8eN1vEy4NGGaY9StRBHPTE6kJk/rGs5eZztP9Ew/MPDzP8Z4D3A2vr3DeOsd1TjPniUev9HxEnAx4ALqY5QAGZHRFdm7m9cQf1+/gRV981sqn27c5znMfo6nwncNUZdx/KafT/rxDz4edUOfm+cHxG7GsZNA/60Hr4VuC8irgb+A/DNzBx9jbdRHTkebB5VUG9rolZB64I6M++NiEWN46I6uXAT1eHfD4ErM/PbVIe7N2XmznrZp1pV1/EWEQuBPwbeStVftz8iHqRq0TwfEbdTtYSeBO7MzGfrRR+j6hZZfpjVH/hjO9x26lkep+r2GHVmw/BjVK2z0zNz39E+13Fso+raWEh15ARVd8X3m1z+WG/z+FlgfUS8Aeih6jZoRuN+WkDVGgX4PeB1wPmZ+UREnAt8ixf2d6P/RlX/T2Xm9oi4hKoR04zHgLMOMf5oX7MzIiIawnoBcEfD9MZ9/Rjw95n59rFWlJkbI+JR4CLg16iCe9TfUXXH/clBi10KfC0z//UI656yjvfJxFuApZn5b4HfB/5XPf61wGvrkxxfGz1Z0iFmUb3xtwJExPuB3obpt1IdXv46L36T/zHwgYg4PyqzIuIXI2L2UW7ndmBZRJxRn8j6w9EJmfk48CVgRUS8PCJOqE+A/dzRP+0Xq1uZtwPLI2J2/Y/ld6kCtBlPAj9+DNvfAnyDqiX4F5n5oyYX/a2ImB8Rc4H/BNxWj59NdbSwq5724cOsYzbwXD3vGcAfHEHpK4H3R8Rb69fljIh4/TG+Zq8AfjsipkfEr1D94xqr1Q7VOYDXRsR76/mnR8TPRERPwzy3Ar8NvJmqj3rUR4E31icn59av+1Kq/u7D7S8d5LgFdUScDLwR+Fzd0vsULxwWTQNeA/w8Vevy041nxSezzNwIrAC+ShU2P0nVFz06/etU/XqvBv6mYfw6qiONT1IdJj9M1b94VNuhCv4vAQ9RtfzuoupDHD1Mfx9V18TGenv/h7EPW4/FUqrn+ggwTPUHvqrJZW8Afrm+wuITR7n9z1Dtlz8db8YGt1Ltt0fqn+vq8R8HTqQ6Uvga8MXDrOOjVCfhngb+GvjLZjeemfdTBdvH6uX/nuqoBI7+Nfs61d/bNmA58MuZuf0Q23+W6qTwu6mOJp6gumJjRsNsg1R/u1/OzG0Ny34XuAB4A1U//C7gj4Bfysy1TdSpWmQe6xHlYVZedX3cmZm9EfFy4DuZ+ZI3UkTcTHUotLp+fDfVSaFvtKy4KS6qS99uzsyF487cISLizVQt+EWZ+Xy762mHiLic6iTnBW3Y9nyqf2ofzsyVx3v7k9lxa1Fn5jPAP9eHWqMX07+hnvwFqkt2iIjTqbpCHjletU0FEXFiRFwc1fXWZ1Aden5+vOU6RURMB5YBn56qId1udffTRcC8+ghbTWpZUEfEINVh+OsiYktE9FP1w/ZHxD9SXYb3znr2vwW2R8RGYAj4g0MdiumoBdUh+E6qro9NwH8Zd6GIm6P6kMnBPze3uN4JU/en7qLqFvh4w/gFh3huz0XEgrYVfIxKfs0y858y87rMfK7dtUwmLe36kCQdOz9CLkmFa8l11KeffnouWrSoFauWpI70wAMPbMvM7rGmtSSoFy1axLp161qxaknqSPUHh8Zk14ckFc6glqTCGdSSVDiDWpIKZ1BLUuEMak0Jg4OD9Pb20tXVRW9vL4ODg+0uSWqaX4qqjjc4OMjAwAArV67kggsuYHh4mP7+fgCWLFnS5uqk8bXkI+R9fX3pddQqRW9vLzfeeCOLFy8+MG5oaIilS5eyfv36NlYmvSAiHsjMvjGnGdTqdF1dXYyMjDB9+vQD4/bu3cvMmTPZv3//YZaUjp/DBbV91Op4PT09DA8Pv2jc8PAwPT09h1hCKotBrY43MDBAf38/Q0ND7N27l6GhIfr7+xkYGGh3aVJTPJmojjd6wnDp0qVs2rSJnp4eli9f7olETRr2UUtSAeyjlqRJzKCWpMIZ1JJUOINakgpnUEtS4QxqSSqcQS1JhTOoJalwBrUkFc6glqTCGdSSVDiDWpIKZ1BLUuEMakkqnEEtSYUzqCWpcE19w0tEbAaeBfYD+w51c2tJ0sQ7kq/iWpyZ21pWiSRpTHZ9SFLhmg3qBL4UEQ9ExFWtLEiS9GLNdn28KTN/EBGvANZGxLcz897GGeoAvwpgwYIFE1ymJE1dTbWoM/MH9e+ngM8D540xzy2Z2ZeZfd3d3RNbpSRNYeMGdUTMiojZo8PALwDrW12YJKnSTNfHK4HPR8To/Ldm5hdbWpUk6YBxgzozHwHecBxqkSSNwcvzJKlwBrUkFc6glqTCGdSSVDiDWpIKZ1BLUuEMakkqnEEtSYUzqCWpcAa1JBXOoJakwhnUklQ4g1qSCmdQS1LhDGpJKpxBLUmFM6glqXAGtSQVzqCWpMIZ1JJUOINakgpnUEtS4QxqSSqcQS1JhTOoJalwBrUkFa7poI6Iroj4VkTc2cqCJEkvdiQt6mXAplYVIkkaW1NBHRHzgV8EPt3aciRJB2u2Rf1x4Brg+UPNEBFXRcS6iFi3devWCSlOktREUEfEO4CnMvOBw82XmbdkZl9m9nV3d09YgZI01TXTon4T8O8jYjPw58BbIuKzLa1KknTAuEGdmddm5vzMXAS8G/hyZr6n5ZVJkgCvo5ak4k07kpkz8x7gnpZUIkkaky1qSSqcQS1JhTOoJalwBrUkFc6glqTCGdSSVDiDWpIKZ1BLUuEMakkqnEEtSYUzqCWpcAa1JBXOoJakwhnUklQ4g1qSCmdQS1LhDGpJKpxBLUmFM6glqXAGtSQVzqCWpMIZ1JJUOINakgpnUEtS4QxqSSqcQS1JhRs3qCNiZkTcHxH/GBEbIuKjx6MwSVJlWhPz7AHekpnPRcR0YDgi/iYzv9bi2iRJNBHUmZnAc/XD6fVPtrIoSdILmuqjjoiuiHgQeApYm5lfH2OeqyJiXUSs27p160TXKUlTVlNBnZn7M/NcYD5wXkT0jjHPLZnZl5l93d3dE12nJE1ZR3TVR2buAu4BLmxJNZKkl2jmqo/uiJhTD58IvA34dqsLkyRVmrnqYx7wmYjoogr22zPzztaWJUka1cxVHw8BP30capEkjcFPJkpS4QxqSSqcQS1JhTOoJalwBrUkFc6glqTCGdSSVDiDWpIKZ1BLUuEMakkqnEEtSYUzqCWpcAa1JBXOoJakwhnUklQ4g1qSCmdQS1LhDGpJKpxBLUmFM6glqXAGtSQVzqCWpMIZ1JJUOINakgpnUEtS4cYN6og4MyKGImJTRGyIiGXHozBJUmVaE/PsA34vM78ZEbOBByJibWZubHFtkiSaaFFn5uOZ+c16+FlgE3BGqwuTJFWOqI86IhYBPw18vRXFSJJequmgjoiTgb8Aficznxlj+lURsS4i1m3dunUia5SkKa2poI6I6VQh/WeZ+ZdjzZOZt2RmX2b2dXd3T2SNkjSlNXPVRwArgU2Z+T9bX5IkqVEzLeo3Ae8F3hIRD9Y/F7e4LklSbdzL8zJzGIjjUIskaQx+MlGSCmdQS1LhDGpJKpxBLUmFM6glqXAGtaaEwcFBent76erqore3l8HBwXaXJDWtmbvnSZPa4OAgAwMDrFy5kgsuuIDh4WH6+/sBWLJkSZurk8YXmTnhK+3r68t169ZN+Hqlo9Hb28uNN97I4sWLD4wbGhpi6dKlrF+/vo2VSS+IiAcys2/MaQa1Ol1XVxcjIyNMnz79wLi9e/cyc+ZM9u/f38bKpBccLqjto1bH6+npYXh4+EXjhoeH6enpaVNF0pExqNXxBgYG6O/vZ2hoiL179zI0NER/fz8DAwPtLk1qiicT1fGWLFnCfffdx0UXXcSePXuYMWMGV155pScSNWnYolbHGxwc5LbbbmPevHlEBPPmzeO2227zEj1NGga1Ot4111xDV1cXq1atYs+ePaxatYquri6uueaadpcmNcWgVsfbsmULa9asYfHixUyfPp3FixezZs0atmzZ0u7SpKYY1JJUOINaHW/+/PlcdtllL7rq47LLLmP+/PntLk1qikGtjnf99dezb98+rrjiCmbOnMkVV1zBvn37uP7669tdmtQUg1odb8mSJdxwww3MmjULgFmzZnHDDTd4eZ4mDT9CLkkF8CPkmvK8zakmMz+ZqI7nbU412dn1oY7nbU41GXibU01p3uZUk4F91JrSvM2pJjuDWh3P25xqshv3ZGJErALeATyVmb2tL0maWKMnDJcuXcqmTZvo6elh+fLlnkjUpDFuH3VEvBl4DljTbFDbRy1JR+aY+qgz815gx4RXJUlqyoT1UUfEVRGxLiLWbd26daJWK0lT3oQFdWbekpl9mdnX3d09UauVJoSfTNRk5lUf6niDg4MsW7aM3bt3k5ns3r2bZcuWGdaaNAxqdTy/ikuT3bhBHRGDwFeB10XElojob31Z0sTxq7g02Y17HXVmerGpJLWRd89Tx5s/fz6XXnopc+bM4dFHH2XhwoXs2rXLr+LSpGEftTreJZdcwjPPPMPIyAgRwcjICM888wyXXHJJu0uTmmJQq+MNDQ1x7bXXctpppwFw2mmnce211zI0NNTmyqTmeJtTdTxvc6rJwNucakrzNqea7DyZqI43MDDAxRdfzMjIyIFxM2fOZNWqVW2sSmqeLWp1vNWrVzMyMsKpp54KwKmnnsrIyAirV69ub2FSkwxqdby1a9dy9dVXs2PHDjKTHTt2cPXVV7N27dp2lyY1xZOJ6ngRwa5duzjllFMOjHv66aeZM2cOrXj/S0fjcCcT7aNWx4sI3vWud/HEE08c+IaXV73qVUREu0uTmmLXhzpeb28vd999N2eddRZPPvkkZ511FnfffTe9vX6znCYHW9TqeM8//zzd3d3ccccdjN4rvbu7m+eff77NlUnNsUWtjrdhwwZ27tzJihUr2L17NytWrGDnzp1s2LCh3aVJTfFkojreCSecwDnnnMPDDz/Mnj17mDFjBmeffTYbN260Va1i+MlETWmZyYYNGzjppJMAOOmkk9iwYYNXfGjSMKg1ZezcufNFv6XJwqCWpMIZ1JJUOINakgpnUEtS4QxqSSqcQS1JhTOoJalwBrUkFc6glqTCGdSSVLimgjoiLoyI70TEwxHxoVYXJUl6wbhBHRFdwE3ARcA5wJKIOKfVhUmSKs20qM8DHs7MRzLzX4E/B97Z2rIkSaOa+YaXM4DHGh5vAc4/eKaIuAq4CmDBggUTUpymmI+cMv48RyE//PLjvk0+8nRr1qspqZmgHusbQF9yI9/MvAW4BaovDjjGujQVtSjcDvcltt6TWpNBM10fW4AzGx7PB37QmnKkiTd37twjGi+Vppmg/gbwmoj4sYh4GfBu4I7WliVNnO3bt78klOfOncv27dvbVJF0ZMbt+sjMfRHxQeBvgS5gVWb6raCaVAxlTWbN9FGTmXcBd7W4FknSGPxkoiQVzqCWpMIZ1JJUOINakgoXrbjgPyK2Ao9O+IqlY3c6sK3dRUhjWJiZ3WNNaElQS6WKiHWZ2dfuOqQjYdeHJBXOoJakwhnUmmpuaXcB0pGyj1qSCmeLWpIKZ1BLUuEMak0JEbEqIp6KiPXtrkU6Uga1porVwIXtLkI6Gga1poTMvBfY0e46pKNhUEtS4QxqSSqcQS1JhTOoJalwBrWmhIgYBL4KvC4itkREf7trkprlR8glqXC2qCWpcAa1JBXOoJakwhnUklQ4g1qSCmdQS1LhDGpJKtz/B1Si1x6xzhyQAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWoAAAEICAYAAAB25L6yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAATO0lEQVR4nO3df3Acd33G8efxWa0cY5LiqA1GiV1aSGXUJFANnjaBRgHS8Ku4DATMjxqswcMM1ZiZUBWPmOK2I9xkpqUeoFATq9QNOX60JNAM+eFMlVKRmFSGhDoRLmlwyNWAZWJjx8ZGtj794zbOWTlZK0Wn++r0fs3c+G53b+9ZnfR477t7d44IAQDStaDeAQAAZ0dRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1EAVtlfYDtsLp3i/TbZvmuZjvsf24HTui8ZGUWNW2b7SdqneOcazvdf2q+udA6iGosaEXLZg3LQp7WECePYo6gZm+0LbX7E9Yvuntj85/qX5+Jf4tu+x3Wf7m5KOSXphNv8Dtr8v6fvZcm+w/YDtQ7bvtX1JxTr32v6Q7e/a/pntL9putr1Y0u2Sltl+MrssO0v+Tba/bPsm20ds/7ftF9veaHu/7cdtX12x/DLbX7P9hO1HbL9v3Lq+ZHt7tq6HbHdk8/5Z0kWS/i3L1FMR4522f2j7gO3enD/65mybj9j+tu1LK3J82Pb/ZvMetv1HZ9n+Ldk2Hra9y/Yr8mxPNv8Zz33FvHW2h20ftH2n7eU5twv1EhE1uUjql7Rf0u6cy18r6WFJD0m6uVa55stFUkHSg5I+LmmxpGZJV0jaJOmmiuVWSApJC7Pb90j6oaSXSFooqSmbv0PS8yQtkvSy7LldlT3OWkl7Jf1yto69ku6XtCy7z7Ck92fzrpRUyrkNmyQdl/QHWZbtkn4gqTfL9T5JP6hY/j8k/X22rZdJGpH0qnHrel2WebOknRX33Svp1VV+Lp/NtvlSSSckteXIPCrpLVnGD2WZm7L5b81+LgskvU3SUUnPz+a9R9JgxbreJWlptu3XSfqxpObJtmei5z6bt1rSI5LasvV+RNK99f595TLJ30LNViy9MvuDnrSoJb1I0nck/Up2+1fr/YOZ6xdJv5sV1cJx0zdp8qL+y3H3CUlXVdz+tKS/GrfMHkm/n13fK+ldFfNukPSZ7PqVmlpR76i4/UZJT0oqZLeXZNnOk3ShpFOSllQsv1nS5yrWdXfFvJWSfl5xe6+qF3VrxbT7Jb09R+bK/wAWSPqRpFdMsPwDkt6UXT+jqKsse1DSpZNtz0TPfTbvdkld4/Idk7S83r+zXCa+1GzoIyK+IemJymm2f8P2HdnLuP+0/VvZrPdJ+lREHMzuu79WueaRCyU9FhEnp3HfxyeZtlzSddmwxyHbh7LHqxzG+HHF9WOSnjONHJL0k4rrP5d0ICJOVdxWtu5lkp6IiCMVyz8m6QVnydScY8x9Ottx+mcVEWOSSlk+2f7jiiGjQ5LaJZ1fbSW2r8uGKH6WLXvuuGUn2p6zPffLJW2pePwnJFln/pyQmNk+MLRV5ZfA37e9SuWXqVdJerEkZeOiBUmbIuKOWc7WaB6XdJHtheP+YI9KOqfi9gVV7lvts28rpz0uqS8i+qaRq1afq7tP0vNsL6ko64sk/V/O+89krgufupIdjG2VtC8bC/6spFdJui8iTtl+QOWiPEM2Hv1n2bIPRcSY7YPVlq1iouf+qXl9EfH56WwY6mPWDibafo6k35P05eyX8x8kPT+bvVDl4Y8rJa2RdKPt82YrW4O6X+WX3H9te3F2MO9ylV9qv9L2RbbPlbRxGuv+rKT3217lssW2X297SY77/kTS0uyxZ0xEPC7pXkmbs229RFKXpLyF9BNJL5yhOL9j+83Z3u0HVR7b3qnyeHGoPCwh2+9VeY+6miWSTmbLLrT955Kem/PxJ3ruJekzkjbafkmW4Vzbb53yFmJWzeZZHwskHYqIyyoubdm8kqSvRsRoRPxA5fHOF81itoaTDQ+8UdJvqnxwsCTpbRGxQ9IXJX1X0i5Jt01j3UMqD1d9UuVx00dUHl/Nc9/vSSpKejR7+T3hWR/TsEblseV9km6R9NFse/PYLOkjWaYPPcscX1X5QOFBSe+W9Obsd/thSX8j6T6V/2P4bUnfnGAdd6o8nvw/Kg/hHFf1IalnmOi5z+bdIul6SV+wfVjSbkmvnfomYjY5onbf8GJ7haTbIqI9u32vpI9HxJdtW9IlEfGg7WskrYmItbbPV/nA4mUR8dOahQOAOaJme9S2iyrvOVxsu2S7S9I7JXXZflDl0/DelC1+p6Sf2n5Y0oCkP6WkAaCspnvUwGRs3y7pFVVmfSwiPjbbefKYi5kxt+Uq6uzA3o0qH/gISesi4r4aZwMAKP/peVsk3RERb7H9Szrz9K5nOP/882PFihXPNhsAzBu7du06EBEt1eZNWtS2n6vyuwzfI0kR8QtJvzjbfVasWKGhoaGpJwWAecr2YxPNy3Mw8YUqn8v5j7a/Y/tGlz9cBwAwC/IU9UKVP7Pj0xHxUpXf2fbh8QvZXm97yPbQyMjIDMcEgPkrT1GXVP4QnW9lt/9F5eI+Q0RsjYiOiOhoaak6zAIAmIZJizoifizpcdsXZ5NepfLHkQIAZkHesz66JX0+O+PjUUnvrV0kAEClXO9MjIgHsmGNSyJi9VMfRwrMFcViUe3t7SoUCmpvb1exWKx3JCA3vv8ODa9YLKq3t1fbtm3TFVdcocHBQXV1dUmS1qxZU+d0wORq8hbyjo6O4DxqpKK9vV2f+MQn1NnZeXrawMCAuru7tXv37jomA55me1dEdFSdR1Gj0RUKBR0/flxNTU2np42Ojqq5uVmnTp06yz2B2XO2ouZbyNHw2traNDg4eMa0wcFBtbW1TXAPIC0UNRpeb2+vurq6NDAwoNHRUQ0MDKirq0u9vb31jgbkwsFENLynDhh2d3dreHhYbW1t6uvr40Ai5gzGqAEgAYxRA8AcRlEDQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEhcri+3tb1X0hFJpySdnOh7vQAAM28q30LeGREHapYEAFAVQx8AkLi8RR2S7rK9y/b6agvYXm97yPbQyMjIzCUEgHkub1FfHhEvk/RaSR+w/crxC0TE1ojoiIiOlpaWGQ0JAPNZrqKOiH3Zv/sl3SLp5bUMBQB42qRFbXux7SVPXZd0taTdtQ4GACjLc9bHr0m6xfZTy98cEXfUNBUA4LRJizoiHpV06SxkAQBUwel5AJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEgcRQ0AiaOoASBxFDUAJI6iBoDEUdQAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJC43EVtu2D7O7Zvq2UgAMCZprJHvUHScK2CAACqy1XUtlslvV7SjbWNAwAYL+8e9d9J6pE0NtECttfbHrI9NDIyMiPhAAA5itr2GyTtj4hdZ1suIrZGREdEdLS0tMxYQACY7/LsUV8u6Q9t75X0BUlX2b6ppqkAAKdNWtQRsTEiWiNihaS3S/r3iHhXzZMBACRxHjUAJG/hVBaOiHsk3VOTJACAqtijBoDEUdQAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEgcRQ0AiaOoASBxFDUAJI6iBoDETVrUtptt32/7QdsP2f6L2QgGAChbmGOZE5KuiognbTdJGrR9e0TsrHE2AIByFHVEhKQns5tN2SVqGQoA8LRcY9S2C7YfkLRf0o6I+FaVZdbbHrI9NDIyMtM5AWDeylXUEXEqIi6T1Crp5bbbqyyzNSI6IqKjpaVlpnMCwLw1pbM+IuKQpHskXVOTNACAZ8hz1keL7fOy64skvVrS92odDABQluesj+dL+ifbBZWL/UsRcVttYwEAnpLnrI/vSnrpLGQBAFTBOxMBIHEUNQAkjqIGgMRR1ACQOIoaABJHUQNA4ihqAEgcRQ0AiaOoASBxFDUAJI6iBoDEUdQAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHGTFrXtC20P2B62/ZDtDbMRDABQtjDHMiclXRcR37a9RNIu2zsi4uEaZwMAKMcedUT8KCK+nV0/ImlY0gtqHQwAUDalMWrbKyS9VNK3qsxbb3vI9tDIyMjMpAMA5C9q28+R9K+SPhgRh8fPj4itEdERER0tLS0zmREA5rVcRW27SeWS/nxEfKW2kQAAlfKc9WFJ2yQNR8Tf1j4SAKBSnj3qyyW9W9JVth/ILq+rcS4AQGbS0/MiYlCSZyELAKAK3pkIAImjqAEgcRQ1ACSOogaAxFHUAJA4ihoAEkdRA0DiKGoASBxFDQCJo6gBIHEUNQAkjqLGvFAsFtXe3q5CoaD29nYVi8V6RwJyo6jR8IrFojZs2KCjR48qInT06FFt2LCBssacQVGj4fX09KhQKKi/v18nTpxQf3+/CoWCenp66h0NyIWiRsMrlUravn27Ojs71dTUpM7OTm3fvl2lUqne0YBcKGoASNykXxwAzHWtra269tprdd555+mxxx7T8uXLdejQIbW2ttY7GpALe9RoeKtXr9bhw4d1/Phx2dbx48d1+PBhrV69ut7RgFwoajS8gYEBbdy4UUuXLpUkLV26VBs3btTAwECdkwH5OCJmfKUdHR0xNDQ04+sFpqNQKOj48eNqamo6PW10dFTNzc06depUHZMBT7O9KyI6qs1jjxoNr62tTYODg2dMGxwcVFtbW50SAVNDUaPh9fb2qqurSwMDAxodHdXAwIC6urrU29tb72hALpz1gYa3Zs0aSVJ3d7eGh4fV1tamvr6+09OB1LFHDQCJY48aDa9YLKq3t1fbtm3TFVdcocHBQXV1dUkSe9WYEybdo7bdb3u/7d2zEQiYaX19fdq2bdsZbyHftm2b+vr66h0NyCXP0MfnJF1T4xxAzQwPD6tUKp3xMaelUknDw8P1jgbkMunQR0R8w/aK2kcBamPZsmXq6enRzTfffHro4x3veIeWLVtW72hALhxMxLxg+6y3gZTNWFHbXm97yPbQyMjITK0WeNb27dun66+/Xt3d3WpublZ3d7euv/567du3r97RgFxmrKgjYmtEdERER0tLy0ytFnjW2tratGfPnjOm7dmzh3cmYs5g6AMNr7OzU5s3b9aBAwc0NjamAwcOaPPmzers7Kx3NCCXPKfnFSXdJ+li2yXbXbWPBcycW2+9VUuWLNGiRYu0YMECLVq0SEuWLNGtt95a72hALnnO+uAdAZjTSqWS7rrrLr3mNa85PW3Hjh26+uqr65gKyI+hDwBIHEWNhtfa2qq1a9ee8el5a9eu5au4MGdQ1Gh4N9xwg06ePKl169apublZ69at08mTJ3XDDTfUOxqQC0WNhrdmzRpt2bJFixcvliQtXrxYW7Zs4QOZMGdQ1ACQOIoaDa9YLGrDhg06evSoIkJHjx7Vhg0bVCwW6x0NyIWiRsPr6elRoVBQf3+/Tpw4of7+fhUKBfX09NQ7GpALRY2GVyqVtH379jM+j3r79u0qlUr1jgbkQlEDQOL4Ki40vNbWVq1evVqjo6MaHR1VU1OTmpqaOI8acwZ71Gh4K1eu1LFjxzQ6OipJGh0d1bFjx7Ry5co6JwPyoajR8O6++27Z1gUXXKAFCxboggsukG3dfffd9Y4G5EJRo+GNjY1p1apVOnjwoMbGxnTw4EGtWrVKY2Nj9Y4G5MIYNeaFnTt3qlAoSJJOnjypnTt31jkRkB971Jg3zjnnHC1YsEDnnHNOvaMAU8IeNeaNI0eOnPEvMFewRw0AiaOoASBxFDUAJI6iBoDEUdQAkDiKGgASR1EDQOIoagBIHEUNAImjqAEgcbmK2vY1tvfYfsT2h2sdCgDwtEmL2nZB0qckvVbSSklrbPOJ6wAwS/J8KNPLJT0SEY9Kku0vSHqTpIdrGQyYjO1ZW0dEPOvHAqYrT1G/QNLjFbdLklaNX8j2eknrJemiiy6akXCYZzadO6XF46PPrVGQKqaYTZt+VpscmJfyFHW1XY5n7F5ExFZJWyWpo6OD3Q9MXY3K7Wx7zewpYy7IczCxJOnCitutkvbVJg4w8yYqY0oac0Weov4vSS+y/eu2f0nS2yV9rbaxgJkVEc+4AHPFpEMfEXHS9p9IulNSQVJ/RDxU82QAAEk5v4orIr4u6es1zgIAqIJ3JgJA4ihqAEgcRQ0AiaOoASBxrsVpSrZHJD024ysGnr3zJR2odwigiuUR0VJtRk2KGkiV7aGI6Kh3DmAqGPoAgMRR1ACQOIoa883WegcApooxagBIHHvUAJA4ihoAEkdRY16w3W97v+3d9c4CTBVFjfnic5KuqXcIYDooaswLEfENSU/UOwcwHRQ1ACSOogaAxFHUAJA4ihoAEkdRY16wXZR0n6SLbZdsd9U7E5AXbyEHgMSxRw0AiaOoASBxFDUAJI6iBoDEUdQAkDiKGgASR1EDQOL+H4iUY2X+xFnzAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWoAAAEICAYAAAB25L6yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAUNklEQVR4nO3dfXBcV33G8eexrFjxS7DdqBDiKC5vqWxRMGjIFKeAgTJJYGDaQhOXAsYixi0I06Z1IGqnaYsy0wlJSUxoKrAJHrCA8FYmBAxtlIAoMcg0YUwUSjBO4phg5cXYkWvLsn/94145a2VtXTla79Hq+5nZ0e49d8/+7mr16Oy5d+86IgQASNe0ahcAADgxghoAEkdQA0DiCGoASBxBDQCJI6gBIHEENcbF9pO2n1ftOk4122H7BeO8zwrbvSf5eK+xvfNk7ovaQ1BjXCJidkRsr3YdlWT7DtvvqXYdwAiCeoqxPb3aNQAYH4K6RtjeYfvDtu+1/YTtT9tuGHkLbfsK249I+rTtabY/ZPsXth+z/UXb8/N+vmX7/aP6vsf2H+fXj04B2H6W7Y22B2w/YPvvbE/L266y/dmSPhbm952e315he7vtfbZ/afvtY2zfCtvft/2vtvfk931lvvwh27ttv6tk/RPVtsJ2r+2P5s/VL21flLd1SvoDSR/Pp3k+XlLG623/PL/PjbZd7FfjdbZ/Y/s+268raXi37f78Odhu+70n6GTk97Uv/x3/0ajnpuz25O3z89fDrrz9ayVtb7J9d/6c/rft3yuwTTjVIoJLDVwk7ZC0TdI5kuZL+r6kj0h6jaRhSf8iaYak0yV9UNJdkhbky/5dUnfezzslfb+k30WS9kiakd8OSS/Ir2+U9B+S5khaKOl/JbXlbVdJ+mxJPwvz+06XNEvSXknn5W1nSVo8xvatyLfj3ZLq8m17UNKN+Ta8QdI+SbML1LZC0iFJl+V9/YWkXZKct98h6T2jHj8k3SpprqQmSQOSLixY819Jqpd0iaTfSJqft79R0vMlWdKrJe2X9LK87TWSdpb09TZJz1U2uLpE0qCkswpuzzckfUHSvLyOV+fLXyZpt6Tz8/u9S9nraEa1X89cRr2WKtaxtCF/EWwruP6fSrpX0k8lbar2EzPZLvkf2OqS2xdL+kX+Bz8kqaGkrV/S60pun5X/oU/Pg21Q0rl5W6ekDSXrhqQX5H/YByUtKml7r6Q78utX6cRBvUfSn0g6veD2rZD085LbL877e3bJssckvbRAbSsk3V/SNjPv6zn57TtUPqgvKLn9RUkfKlDz0cDMl/1Q0juOs/7XJK3Jr79GJUFdZt27Jb1lrO3Jf7dHJM0r08e/SfrnUct+pjzIuaRzqeTUx82SLiyyou0XSvqwpKURsVjZiA/j91DJ9QeUjcAkaSAiDpS0nSvpq/nb3T3KgvuwstDbp2wEdmm+7qWSPlfmsc6UdFr+OKWPefZYRUbEoLJR4WpJv7L9Ddu/O9b9JP265Pr/5X2NXja7YG2PlNSzP786e4zHf6Tk+v4C60vSw5EnYEkdz5Uk2xfZvsv24/nv4eK89qex/c6SKYo9klpGrXu87TlH0uMR8USZbs+VdPlIn3m/5+ip1w0SUbGgjojvSnq8dJnt5+dzoFttf6/kj/MySTeOvJgiYnel6qpx55Rcb1I2mpOy0VWphyRdFBFzSy4NEfFw3t4tabnt31c2VdJT5rEeVTYKP3fUY470MahsZDfiOaV3jojNEfGHykZ890n6ZJENLGis2sYykaeUPHvUXHaTpF22Z0j6sqSPKvsHOVfSbcqmQY5h+1xlz8/7Jf1Wvu62cuuW8ZCk+bbnHqetc9TrYGZEdI9nA1F5p3pnYpek9oh4uaS/kfSJfPmLJL0o31l0l+1CI3E8zftsL8h3DF6pbF6ynJskdeYBINuNtt9S0n6bspD7J0lfiIgjozuIiMPK3v532p6T9/XXkkZ2IN4t6VW2m2w/S9k7JuWP92zbb7Y9S9kUxZPKRvQTokBtY/m1pIk6Vvy3JX3Adr3tt0lqVvb8nqZsbn1A0nC+8+8Nx+ljlrJ/HgNSthNS2Yh6TBHxK0nflPQJ2/PyOl6VN39S0mrb5zszy/Ybbc85uU1FpZyyoLY9W9IrJd1i+25lO7DOypunS3qhsnm55ZI+dZwRAE5sk6RvS9qeXz5ynPWul/R1Sd+2vU/ZjsXzRxoj4qCkr0h6fd7n8bQrGzlvl9Sbr7sh7+M7yv5R/ETSVmU74kZMk3S5shH/48p2pP1l8c0s5Li1FXC9pLfmR0jc8Azr2KLstf2osvn+t0bEY/kU0weU/UN5QtKfKfudPE1E3CvpWkk/UPZP5MXKdhYX9Q5l7zDuU7bf6IN5v33K3s1+PK/hfmXz3UiMj50+m+DO7YWSbo2IFttnSPpZRJxVZr2bJN0VETfnt/9L2Y6aH1WsuBpje4eyHWD/We1aAEysUzaijoi9kn6Zv/1T/lbrJXnz1yQty5efqWwqpKY//QYARRUKattzbX8pP2C/P9/JNNZ9upW9VTvP2Qcu2iS9XVKb7XuUHYY3Mi+6WdJjtu9VtuPqbyPisZPZIExetm/KP2Qy+nJTtWs7nslYMyafQlMftj8j6XsR8Snbp0maGRF7Kl4dAGDsoM7nlu+R9Lyo5IQ2AKCsIifoeZ6yw4I+nc8pb1X26anB493hzDPPjIULF05MhQAwBWzduvXRiGgs11ZkRN2q7PCtpRGxxfb1kvZGxN+PWm+VpFWS1NTU9PIHHnjg6Z0BAMqyvTUiWsu1FdmZuFPZOQe25Le/pOxkLseIiK6IaI2I1sbGsv8UAAAnYcygjohHJD1k+7x80euUnTwJAHAKFD2JfLukz+VHfGxXdqpJAMApUCioI+JuSWXnTgAAlcU3vABA4ghqTAnd3d1qaWlRXV2dWlpa1N3NmTwxefBFp6h53d3d6ujo0Pr163XBBReot7dXbW1tkqTly5dXuTpgbBU5e15ra2v09fVNeL/AyWhpadG6deu0bNmyo8t6enrU3t6ubdu2VbEy4CknOo6aoEbNq6ur04EDB1RfX3902aFDh9TQ0KDDhyfs+wqAZ+SZfuAFmNSam5vV29t7zLLe3l41NzdXqSJgfAhq1LyOjg61tbWpp6dHhw4dUk9Pj9ra2tTR0VHt0oBC2JmImjeyw7C9vV39/f1qbm5WZ2cnOxIxaTBHDQAJYI4aACYxghoAEkdQA0DiCGoASBxBDQCJI6gBIHEENQAkjqAGgMQR1ACQOIIaABJHUANA4ghqAEgcQQ0AiSOoASBxBDUAJI6gBoDEEdQAkLhCX8Vle4ekfZIOSxo+3rcQAAAm3ni+M3FZRDxasUoAAGUx9QEAiSsa1CHp27a32l5VbgXbq2z32e4bGBiYuAoBYIorGtRLI+Jlki6S9D7brxq9QkR0RURrRLQ2NjZOaJEAMJUVCuqI2JX/3C3pq5JeUcmiAABPGTOobc+yPWfkuqQ3SNpW6cIAAJkiR308W9JXbY+svykivlXRqgAAR40Z1BGxXdJLTkEtAIAyODwPABJHUANA4ghqAEgcQQ0AiSOoASBxBDUAJI6gBoDEEdQAkDiCGgASR1ADQOIIagBIHEENAIkjqAEgcQQ1ACSOoAaAxBHUAJA4ghoAEkdQA0DiCGoASBxBDQCJI6gBIHEENQAkjqAGgMQR1ACQOIIaABJXOKht19n+H9u3VrIgAMCxxjOiXiOpv1KFAADKKxTUthdIeqOkT1W2HADAaEVH1B+TtFbSkQrWAgAoY8ygtv0mSbsjYusY662y3We7b2BgYMIKBICprsiIeqmkN9veIenzkl5r+7OjV4qIrohojYjWxsbGCS4TAKauMYM6Ij4cEQsiYqGkSyXdHhF/XvHKAACSOI4aAJI3fTwrR8Qdku6oSCUAgLIYUQNA4ghqAEgcQQ0AiSOoASBxBDUAJI6gBoDEEdQAkDiCGgASR1ADQOIIagBIHEENAIkjqAEgcQQ1ACSOoAaAxBHUAJA4ghoAEkdQA0DiCGoASBxBDQCJI6gBIHEENQAkjqAGgMQR1ACQOIIaABJHUANA4ghqAEjcmEFtu8H2D23fY/untv/xVBQGAMhML7DOQUmvjYgnbddL6rX9zYi4q8K1AQBUIKgjIiQ9md+szy9RyaIAAE8pNEdtu8723ZJ2S/pORGwps84q2322+wYGBia6TgCYsgoFdUQcjoiXSlog6RW2W8qs0xURrRHR2tjYONF1AsCUNa6jPiJij6Q7JF1YkWoAAE9T5KiPRttz8+unS3q9pPsqXRgAIFPkqI+zJH3Gdp2yYP9iRNxa2bIAACOKHPXxE0lLTkEtAIAy+GQiACSOoAaAxBHUAJA4ghoAEkdQA0DiCGoASBxBDQCJI6gBIHEENQAkjqAGgMQR1ACQOIIaABJHUANA4ghqAEgcQQ0AiSOoASBxBDUAJI6gBoDEEdQAkDiCGgASR1ADQOIIagBIHEENAIkjqAEgcQQ1ACSOoAaAxI0Z1LbPsd1ju9/2T22vORWFAQAy0wusMyzp8oj4se05krba/k5E3Fvh2gAAKjCijohfRcSP8+v7JPVLOrvShQEAMuOao7a9UNISSVvKtK2y3We7b2BgYGKqAwAUD2rbsyV9WdIHI2Lv6PaI6IqI1ohobWxsnMgaAWBKKxTUtuuVhfTnIuIrlS0JAFCqyFEflrReUn9EXFf5kgAApYqMqJdKeoek19q+O79cXOG6AAC5MQ/Pi4heST4FtQAAyuCTiQCQOIIaABJHUANA4ghqAEgcQQ0AiSOoMSV0d3erpaVFdXV1amlpUXd3d7VLAgorcvY8YFLr7u5WR0eH1q9frwsuuEC9vb1qa2uTJC1fvrzK1QFjc0RMeKetra3R19c34f0CJ6OlpUXr1q3TsmXLji7r6elRe3u7tm3bVsXKgKfY3hoRrWXbCGrUurq6Oh04cED19fVHlx06dEgNDQ06fPhwFSsDnnKioGaOGjWvublZvb29xyzr7e1Vc3NzlSoCxoc5atS8jo4OXXLJJZo1a5YefPBBNTU1aXBwUNdff321SwMKIagxJRw8eFB79uzRkSNH9PDDD+v000+vdklAYUx9oOatXbtWM2fO1ObNmzU0NKTNmzdr5syZWrt2bbVLAwohqFHzdu7cqY0bN2rZsmWqr6/XsmXLtHHjRu3cubPapQGFENSYEm6//fZjPvBy++23V7skoDCCGjVv/vz5uuaaa7Ry5Urt27dPK1eu1DXXXKP58+dXuzSgEIIaNW/mzJmaM2eO1q1bd8zPmTNnVrs0oBCCGjVv165duuGGGzRr1ixJ0qxZs3TDDTdo165dVa4MKIbD81DzmpubtWDBgmM+Lt7T08MHXjBpMKJGzevo6FBbW5t6enp06NAh9fT0qK2tTR0dHdUuDSiEETVq3sgZ8trb29Xf36/m5mZ1dnZy5jxMGpyUCQASwEmZAGASI6gBIHEENQAkbsygtr3B9m7bfBUGAFRBkRH1zZIurHAdQEXx5baYzMY8PC8ivmt7YeVLASqDL7fFZDdhc9S2V9nus903MDAwUd0Cz1hnZ6fWr19/zGlO169fr87OzmqXBhQyYUEdEV0R0RoRrY2NjRPVLfCM9ff365ZbblFDQ4Nsq6GhQbfccov6+/urXRpQCEd9oObNnTtXXV1duvrqqzU4OKirr75aXV1dmjt3brVLAwohqFHz9u7dqzPOOENLlixRfX29lixZojPOOEN79+6tdmlAIUUOz+uW9ANJ59neabut8mUBE2d4eFjXXnut2tvb1dDQoPb2dl177bUaHh6udmlAIUWO+mC3OCa1GTNm6IknnjjmNKfXXXedZsyYUcWqgOI4ex5q3mWXXaYrrrhCkrR69WrddNNNuuKKK7R69eoqVwYUQ1Cj5q1bt06SdOWVV+ryyy/XjBkztHr16qPLgdRxmlMASACnOcWUx0fIMZkR1Kh53d3dWrNmjQYHBxURGhwc1Jo1awhrTBoENWre2rVrNTQ0JEmyLUkaGhrS2rVrq1kWUBhBjZq3c+fOo9dL98mULgdSRlBjSpg2bZo2bNiggwcPasOGDZo2jZc+Jg9erZgSRqY8jncbSBnHUWNKGB4e1sqVK/Xggw+qqamJj49jUmFEjZq3YMECHThwQDt27NCRI0e0Y8cOHThwQAsWLKh2aUAhBDVq3qJFizQ0NKR58+bJtubNm6ehoSEtWrSo2qUBhRDUqHl33nmnli5dqv379ysitH//fi1dulR33nlntUsDCmGOGjXv4MGD2rJly9F56dG3gdQxosaUMDw8rNmzZ0uSZs+eTUhjUiGoMWU8+eSTx/wEJguCGgASR1ADQOIIagBIHEENAIkjqAEgcQQ1ACSOoAaAxBHUAJA4ghoAEkdQA0DiCgW17Qtt/8z2/bY/VOmiAABPGTOobddJulHSRZIWSVpumxP5AsApUuQ0p6+QdH9EbJck25+X9BZJ91ayMGAsE/G9h0X7KP32cuBUKxLUZ0t6qOT2Tknnj17J9ipJqySpqalpQorDFHPVs8a1evzDGRUqpIxx1qarflOZOjAlFQnqckOOpw0vIqJLUpcktba2MvzA+FUo3E40amakjMmgyM7EnZLOKbm9QNKuypQDTLxNmzaNazmQmiJB/SNJL7T9O7ZPk3SppK9Xtixg4ixfvlybNm3S4sWLNW3aNC1evFibNm3S8uXLq10aUIiLvPWzfbGkj0mqk7QhIjpPtH5ra2v09fVNTIUAMAXY3hoRreXaCn25bUTcJum2Ca0KAFAIn0wEgMQR1ACQOIIaABJHUANA4god9THuTu0BSQ9MeMfAM3empEerXQRQxrkR0ViuoSJBDaTKdt/xDoECUsXUBwAkjqAGgMQR1JhquqpdADBezFEDQOIYUQNA4ghqAEgcQY0pwfYG27ttb6t2LcB4EdSYKm6WdGG1iwBOBkGNKSEivivp8WrXAZwMghoAEkdQA0DiCGoASBxBDQCJI6gxJdjulvQDSefZ3mm7rdo1AUXxEXIASBwjagBIHEENAIkjqAEgcQQ1ACSOoAaAxBHUAJA4ghoAEvf/onO2PU1W+hoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "for i in data.columns:\n",
    "    if data[i].dtype != 'object':\n",
    "        fig, axs = plt.subplots()\n",
    "        axs.boxplot(data[i])\n",
    "        axs.set_title(i)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b> Here I have not dealt with the outliers because the data is of money and such variations do exist in real banks so reducing them will effect the generality of my model so I have left it as it is. <b/>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# data.loc[data['vintage']>5500,'vintage']=np.mean(data['vintage'])\n",
    "# data['vintage'].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['days_since_last_transaction']>90,'days_since_last_transaction']=np.mean(data['days_since_last_transaction'])\n",
    "# data['days_since_last_transaction'].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# #np.power(data['current_balance'],1/3).plot.hist()\n",
    "# data['current_balance']=np.log(data['current_balance'])\n",
    "# data['current_balance'].plot.box()\n",
    "# # data['current_balance'].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# data.loc[data['current_balance']<0.4,'current_balance']=np.mean(data['current_balance'])\n",
    "# data['current_balance'].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data.loc[data['']>90,'']=np.mean(data[''])\n",
    "# data[''].plot.box()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using One-Hot Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "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>customer_id</th>\n",
       "      <th>vintage</th>\n",
       "      <th>age</th>\n",
       "      <th>branch_code</th>\n",
       "      <th>days_since_last_transaction</th>\n",
       "      <th>current_balance</th>\n",
       "      <th>previous_month_end_balance</th>\n",
       "      <th>average_monthly_balance_prevQ</th>\n",
       "      <th>average_monthly_balance_prevQ2</th>\n",
       "      <th>current_month_credit</th>\n",
       "      <th>...</th>\n",
       "      <th>city_1643.0</th>\n",
       "      <th>city_1644.0</th>\n",
       "      <th>city_1645.0</th>\n",
       "      <th>city_1646.0</th>\n",
       "      <th>city_1647.0</th>\n",
       "      <th>city_1648.0</th>\n",
       "      <th>city_1649.0</th>\n",
       "      <th>customer_nw_category_1</th>\n",
       "      <th>customer_nw_category_2</th>\n",
       "      <th>customer_nw_category_3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>3135</td>\n",
       "      <td>66</td>\n",
       "      <td>755</td>\n",
       "      <td>224.0</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1458.71</td>\n",
       "      <td>1449.07</td>\n",
       "      <td>0.20</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>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>310</td>\n",
       "      <td>35</td>\n",
       "      <td>3214</td>\n",
       "      <td>60.0</td>\n",
       "      <td>5390.37</td>\n",
       "      <td>8704.66</td>\n",
       "      <td>7799.26</td>\n",
       "      <td>12419.41</td>\n",
       "      <td>0.56</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>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4</td>\n",
       "      <td>2356</td>\n",
       "      <td>31</td>\n",
       "      <td>41</td>\n",
       "      <td>30.0</td>\n",
       "      <td>3913.16</td>\n",
       "      <td>5815.29</td>\n",
       "      <td>4910.17</td>\n",
       "      <td>2815.94</td>\n",
       "      <td>0.61</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>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5</td>\n",
       "      <td>478</td>\n",
       "      <td>90</td>\n",
       "      <td>582</td>\n",
       "      <td>147.0</td>\n",
       "      <td>2291.91</td>\n",
       "      <td>2291.91</td>\n",
       "      <td>2084.54</td>\n",
       "      <td>1006.54</td>\n",
       "      <td>0.47</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>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6</td>\n",
       "      <td>2531</td>\n",
       "      <td>42</td>\n",
       "      <td>388</td>\n",
       "      <td>58.0</td>\n",
       "      <td>927.72</td>\n",
       "      <td>1401.72</td>\n",
       "      <td>1643.31</td>\n",
       "      <td>1871.12</td>\n",
       "      <td>0.33</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>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 1645 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  vintage  age  branch_code  days_since_last_transaction  \\\n",
       "0            1     3135   66          755                        224.0   \n",
       "1            2      310   35         3214                         60.0   \n",
       "2            4     2356   31           41                         30.0   \n",
       "3            5      478   90          582                        147.0   \n",
       "4            6     2531   42          388                         58.0   \n",
       "\n",
       "   current_balance  previous_month_end_balance  average_monthly_balance_prevQ  \\\n",
       "0          1458.71                     1458.71                        1458.71   \n",
       "1          5390.37                     8704.66                        7799.26   \n",
       "2          3913.16                     5815.29                        4910.17   \n",
       "3          2291.91                     2291.91                        2084.54   \n",
       "4           927.72                     1401.72                        1643.31   \n",
       "\n",
       "   average_monthly_balance_prevQ2  current_month_credit  ...  city_1643.0  \\\n",
       "0                         1449.07                  0.20  ...            0   \n",
       "1                        12419.41                  0.56  ...            0   \n",
       "2                         2815.94                  0.61  ...            0   \n",
       "3                         1006.54                  0.47  ...            0   \n",
       "4                         1871.12                  0.33  ...            0   \n",
       "\n",
       "   city_1644.0  city_1645.0  city_1646.0  city_1647.0  city_1648.0  \\\n",
       "0            0            0            0            0            0   \n",
       "1            0            0            0            0            0   \n",
       "2            0            0            0            0            0   \n",
       "3            0            0            0            0            0   \n",
       "4            0            0            0            0            0   \n",
       "\n",
       "   city_1649.0  customer_nw_category_1  customer_nw_category_2  \\\n",
       "0            0                       0                       1   \n",
       "1            0                       0                       1   \n",
       "2            0                       0                       1   \n",
       "3            0                       0                       1   \n",
       "4            0                       0                       0   \n",
       "\n",
       "   customer_nw_category_3  \n",
       "0                       0  \n",
       "1                       0  \n",
       "2                       0  \n",
       "3                       0  \n",
       "4                       1  \n",
       "\n",
       "[5 rows x 1645 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data=pd.get_dummies(data)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Separating independent and dependent variables "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((28382, 1644), (28382,))"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = data.drop(['churn'],axis=1)  # independent variable\n",
    "y = data['churn']  # dependent variable\n",
    "x.shape, y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b> Using Standard Scaler to scale the data </b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "scaler=StandardScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "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>customer_id</th>\n",
       "      <th>vintage</th>\n",
       "      <th>age</th>\n",
       "      <th>branch_code</th>\n",
       "      <th>days_since_last_transaction</th>\n",
       "      <th>current_balance</th>\n",
       "      <th>previous_month_end_balance</th>\n",
       "      <th>average_monthly_balance_prevQ</th>\n",
       "      <th>average_monthly_balance_prevQ2</th>\n",
       "      <th>current_month_credit</th>\n",
       "      <th>...</th>\n",
       "      <th>city_1643.0</th>\n",
       "      <th>city_1644.0</th>\n",
       "      <th>city_1645.0</th>\n",
       "      <th>city_1646.0</th>\n",
       "      <th>city_1647.0</th>\n",
       "      <th>city_1648.0</th>\n",
       "      <th>city_1649.0</th>\n",
       "      <th>customer_nw_category_1</th>\n",
       "      <th>customer_nw_category_2</th>\n",
       "      <th>customer_nw_category_3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.731304</td>\n",
       "      <td>0.478644</td>\n",
       "      <td>0.999147</td>\n",
       "      <td>-0.182318</td>\n",
       "      <td>1.927027</td>\n",
       "      <td>-0.139017</td>\n",
       "      <td>-0.141953</td>\n",
       "      <td>-0.144709</td>\n",
       "      <td>-0.127317</td>\n",
       "      <td>-0.044545</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.043254</td>\n",
       "      <td>-0.018774</td>\n",
       "      <td>-0.015707</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.014541</td>\n",
       "      <td>-0.028479</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.38784</td>\n",
       "      <td>0.974396</td>\n",
       "      <td>-0.743959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.731190</td>\n",
       "      <td>-1.275909</td>\n",
       "      <td>-0.741756</td>\n",
       "      <td>2.439824</td>\n",
       "      <td>-0.066312</td>\n",
       "      <td>-0.046720</td>\n",
       "      <td>0.028425</td>\n",
       "      <td>0.007249</td>\n",
       "      <td>0.118793</td>\n",
       "      <td>-0.044540</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.043254</td>\n",
       "      <td>-0.018774</td>\n",
       "      <td>-0.015707</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.014541</td>\n",
       "      <td>-0.028479</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.38784</td>\n",
       "      <td>0.974396</td>\n",
       "      <td>-0.743959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-1.730961</td>\n",
       "      <td>-0.005178</td>\n",
       "      <td>-0.966389</td>\n",
       "      <td>-0.943689</td>\n",
       "      <td>-0.430947</td>\n",
       "      <td>-0.081398</td>\n",
       "      <td>-0.039514</td>\n",
       "      <td>-0.061991</td>\n",
       "      <td>-0.096652</td>\n",
       "      <td>-0.044539</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.043254</td>\n",
       "      <td>-0.018774</td>\n",
       "      <td>-0.015707</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.014541</td>\n",
       "      <td>-0.028479</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.38784</td>\n",
       "      <td>0.974396</td>\n",
       "      <td>-0.743959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-1.730847</td>\n",
       "      <td>-1.171568</td>\n",
       "      <td>2.346943</td>\n",
       "      <td>-0.366796</td>\n",
       "      <td>0.991130</td>\n",
       "      <td>-0.119457</td>\n",
       "      <td>-0.122361</td>\n",
       "      <td>-0.129711</td>\n",
       "      <td>-0.137244</td>\n",
       "      <td>-0.044541</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.043254</td>\n",
       "      <td>-0.018774</td>\n",
       "      <td>-0.015707</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.014541</td>\n",
       "      <td>-0.028479</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.38784</td>\n",
       "      <td>0.974396</td>\n",
       "      <td>-0.743959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.730733</td>\n",
       "      <td>0.103512</td>\n",
       "      <td>-0.348649</td>\n",
       "      <td>-0.573667</td>\n",
       "      <td>-0.090621</td>\n",
       "      <td>-0.151482</td>\n",
       "      <td>-0.143293</td>\n",
       "      <td>-0.140285</td>\n",
       "      <td>-0.117848</td>\n",
       "      <td>-0.044543</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.043254</td>\n",
       "      <td>-0.018774</td>\n",
       "      <td>-0.015707</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.014541</td>\n",
       "      <td>-0.028479</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.38784</td>\n",
       "      <td>-1.026277</td>\n",
       "      <td>1.344159</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 1645 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id   vintage       age  branch_code  days_since_last_transaction  \\\n",
       "0    -1.731304  0.478644  0.999147    -0.182318                     1.927027   \n",
       "1    -1.731190 -1.275909 -0.741756     2.439824                    -0.066312   \n",
       "2    -1.730961 -0.005178 -0.966389    -0.943689                    -0.430947   \n",
       "3    -1.730847 -1.171568  2.346943    -0.366796                     0.991130   \n",
       "4    -1.730733  0.103512 -0.348649    -0.573667                    -0.090621   \n",
       "\n",
       "   current_balance  previous_month_end_balance  average_monthly_balance_prevQ  \\\n",
       "0        -0.139017                   -0.141953                      -0.144709   \n",
       "1        -0.046720                    0.028425                       0.007249   \n",
       "2        -0.081398                   -0.039514                      -0.061991   \n",
       "3        -0.119457                   -0.122361                      -0.129711   \n",
       "4        -0.151482                   -0.143293                      -0.140285   \n",
       "\n",
       "   average_monthly_balance_prevQ2  current_month_credit  ...  city_1643.0  \\\n",
       "0                       -0.127317             -0.044545  ...    -0.043254   \n",
       "1                        0.118793             -0.044540  ...    -0.043254   \n",
       "2                       -0.096652             -0.044539  ...    -0.043254   \n",
       "3                       -0.137244             -0.044541  ...    -0.043254   \n",
       "4                       -0.117848             -0.044543  ...    -0.043254   \n",
       "\n",
       "   city_1644.0  city_1645.0  city_1646.0  city_1647.0  city_1648.0  \\\n",
       "0    -0.018774    -0.015707    -0.005936    -0.014541    -0.028479   \n",
       "1    -0.018774    -0.015707    -0.005936    -0.014541    -0.028479   \n",
       "2    -0.018774    -0.015707    -0.005936    -0.014541    -0.028479   \n",
       "3    -0.018774    -0.015707    -0.005936    -0.014541    -0.028479   \n",
       "4    -0.018774    -0.015707    -0.005936    -0.014541    -0.028479   \n",
       "\n",
       "   city_1649.0  customer_nw_category_1  customer_nw_category_2  \\\n",
       "0    -0.005936                -0.38784                0.974396   \n",
       "1    -0.005936                -0.38784                0.974396   \n",
       "2    -0.005936                -0.38784                0.974396   \n",
       "3    -0.005936                -0.38784                0.974396   \n",
       "4    -0.005936                -0.38784               -1.026277   \n",
       "\n",
       "   customer_nw_category_3  \n",
       "0               -0.743959  \n",
       "1               -0.743959  \n",
       "2               -0.743959  \n",
       "3               -0.743959  \n",
       "4                1.344159  \n",
       "\n",
       "[5 rows x 1645 columns]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_data=scaler.fit_transform(data)\n",
    "scaled_data=pd.DataFrame(scaled_data,columns=data.columns)\n",
    "scaled_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((28382, 1644), (28382,))"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_x = scaled_data.drop(['churn'],axis=1)\n",
    "scaled_y = scaled_data['churn']\n",
    "scaled_x.shape, scaled_y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "train_x,test_x,train_y,test_y = train_test_split(x,y, random_state = 56,test_size=0.30, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaled_train_x,scaled_test_x,scaled_train_y,scaled_test_y = train_test_split(scaled_x,scaled_y, random_state = 56,\n",
    "                                                                             test_size=0.30, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0    0.814668\n",
       " 1    0.185332\n",
       " Name: churn, dtype: float64,\n",
       " 0    0.81468\n",
       " 1    0.18532\n",
       " Name: churn, dtype: float64)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_y.value_counts(normalize=True),test_y.value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression as LogReg\n",
    "from sklearn.metrics import f1_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    19867.000000\n",
       "mean         0.185332\n",
       "std          0.388577\n",
       "min          0.000000\n",
       "25%          0.000000\n",
       "50%          0.000000\n",
       "75%          0.000000\n",
       "max          1.000000\n",
       "Name: churn, dtype: float64"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_y.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='auto', n_jobs=-1, penalty='l2',\n",
       "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating instance of Logistic Regresssion\n",
    "LR = LogReg(n_jobs=-1)\n",
    "\n",
    "# Fitting the model\n",
    "LR.fit(train_x, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_predict = LR.predict(train_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_predict = LR.predict(test_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training f1_score 0.1466275659824047\n"
     ]
    }
   ],
   "source": [
    "k = f1_score(train_predict, train_y)\n",
    "print('Training f1_score', k )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test f1_score     0.15584415584415584\n"
     ]
    }
   ],
   "source": [
    "k = f1_score(test_predict, test_y)\n",
    "print('Test f1_score    ', k )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "le=LabelEncoder()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((19867,), (8515,))"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_train_y=le.fit_transform(scaled_train_y)\n",
    "scaled_test_y=le.fit_transform(scaled_test_y)\n",
    "scaled_train_y.shape, scaled_test_y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# K- Nearest Neighbours (KNN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "#importing KNN classifier\n",
    "from sklearn.neighbors import KNeighborsClassifier as KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "                     metric_params=None, n_jobs=-1, n_neighbors=10, p=2,\n",
       "                     weights='uniform')"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating instance of KNN\n",
    "clf = KNN(n_neighbors = 10,n_jobs=-1)\n",
    "\n",
    "# Fitting the model\n",
    "clf.fit(scaled_train_x, scaled_train_y)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test F1 Score     0.016019716574245228\n"
     ]
    }
   ],
   "source": [
    "# Predicting over the Train Set and calculating F1\n",
    "test_predict = clf.predict(scaled_test_x)\n",
    "k = f1_score(test_predict, scaled_test_y)\n",
    "print('Test F1 Score    ', k )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test F1 Score     0.41642228739002934\n"
     ]
    }
   ],
   "source": [
    "# Creating instance of KNN\n",
    "clf = KNN(n_neighbors = 1)\n",
    "\n",
    "# Fitting the model\n",
    "clf.fit(train_x, train_y)\n",
    "\n",
    "# Predicting over the Train Set and calculating F1\n",
    "test_predict = clf.predict(test_x)\n",
    "k = f1_score(test_predict, test_y)\n",
    "print('Test F1 Score    ', k )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test F1 Score     1.0\n"
     ]
    }
   ],
   "source": [
    "train_predict = clf.predict(train_x)\n",
    "k = f1_score(train_predict, train_y)\n",
    "print('Test F1 Score    ', k )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Elbow(K):\n",
    "    #initiating empty list\n",
    "    test_error = []\n",
    "   \n",
    "    #training model for evey value of K\n",
    "    for i in K:\n",
    "        #Instance oh KNN\n",
    "        clf = KNN(n_neighbors = i)\n",
    "        clf.fit(train_x, train_y)\n",
    "        # Appending F1 scores to empty list claculated using the predictions\n",
    "        tmp = clf.predict(test_x)\n",
    "        tmp = f1_score(tmp,test_y)\n",
    "        error = 1-tmp\n",
    "        test_error.append(error)\n",
    "    \n",
    "    return test_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "61WGHNM_Cxn2"
   },
   "outputs": [],
   "source": [
    "#Defining K range\n",
    "k = range(6, 20, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SNBDTcSf9xXW"
   },
   "outputs": [],
   "source": [
    "# calling above defined function\n",
    "test = Elbow(k)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Elbow curve</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2854,
     "status": "ok",
     "timestamp": 1555058107314,
     "user": {
      "displayName": "Sharoon Saxena",
      "photoUrl": "",
      "userId": "14774175216384036942"
     },
     "user_tz": -330
    },
    "id": "6iA6n55NDKJf",
    "outputId": "bb6700af-e76f-433f-b81e-92537cb3a60c",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Elbow Curve for test')"
      ]
     },
     "execution_count": 52,
     "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": [
    "# plotting the Curves\n",
    "plt.plot(k, test)\n",
    "plt.xlabel('K Neighbors')\n",
    "plt.ylabel('Test error')\n",
    "plt.title('Elbow Curve for test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "#importing decision tree classifier \n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "dt_model = DecisionTreeClassifier(random_state=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(ccp_alpha=0.0, class_weight=None, criterion='gini',\n",
       "                       max_depth=None, max_features=None, max_leaf_nodes=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, presort='deprecated',\n",
       "                       random_state=10, splitter='best')"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#fitting the model\n",
    "dt_model.fit(train_x, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#checking the training score\n",
    "dt_model.score(train_x, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7963593658250147"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#checking the test score\n",
    "dt_model.score(test_x, test_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_accuracy = []\n",
    "validation_accuracy = []\n",
    "for depth in range(1,10):\n",
    "    dt_model = DecisionTreeClassifier(max_depth=depth, random_state=10)\n",
    "    dt_model.fit(train_x, train_y)\n",
    "    train_accuracy.append(dt_model.score(train_x, train_y))\n",
    "    validation_accuracy.append(dt_model.score(test_x, test_y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "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>max_depth</th>\n",
       "      <th>train_acc</th>\n",
       "      <th>valid_acc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.846227</td>\n",
       "      <td>0.846154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>0.846227</td>\n",
       "      <td>0.846154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>0.851764</td>\n",
       "      <td>0.851204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0.858660</td>\n",
       "      <td>0.847563</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>0.864851</td>\n",
       "      <td>0.849442</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   max_depth  train_acc  valid_acc\n",
       "0          1   0.846227   0.846154\n",
       "1          2   0.846227   0.846154\n",
       "2          3   0.851764   0.851204\n",
       "3          4   0.858660   0.847563\n",
       "4          5   0.864851   0.849442"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame = pd.DataFrame({'max_depth':range(1,10), 'train_acc':train_accuracy, 'valid_acc':validation_accuracy})\n",
    "frame.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b> Plotting performance vs depth of tree curve to find the best value of max_depth </b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x21f5ebcabe0>"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,6))\n",
    "plt.plot(frame['max_depth'], frame['train_acc'], marker='o')\n",
    "plt.plot(frame['max_depth'], frame['valid_acc'], marker='o')\n",
    "plt.xlabel('Depth of tree')\n",
    "plt.ylabel('performance')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b> Putting max_depth=3 as it gives the best fit model as per the curve above </b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "dt_model = DecisionTreeClassifier(max_depth=3, random_state=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(ccp_alpha=0.0, class_weight=None, criterion='gini',\n",
       "                       max_depth=3, max_features=None, max_leaf_nodes=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, presort='deprecated',\n",
       "                       random_state=10, splitter='best')"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#fitting the model\n",
    "dt_model.fit(train_x, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.851764232143756"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#checking the training score\n",
    "dt_model.score(train_x, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8512037580739871"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#checking the test score\n",
    "dt_model.score(test_x, test_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Importing random forest classifier \n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "#creating a random forest instance\n",
    "RFclf = RandomForestClassifier(random_state=96,n_jobs=-1,criterion='entropy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,\n",
       "                       criterion='entropy', max_depth=None, max_features='auto',\n",
       "                       max_leaf_nodes=None, max_samples=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "                       n_jobs=-1, oob_score=False, random_state=96, verbose=0,\n",
       "                       warm_start=False)"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#train the model\n",
    "RFclf.fit(scaled_train_x,scaled_train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#score on training data\n",
    "RFclf.score(scaled_train_x, scaled_train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8563711098062243"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#score on test data\n",
    "RFclf.score(scaled_test_x, scaled_test_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Final Conclusion:\n",
    "### I have made models using Logistic Regression, KNN, Decision Tree and Random Forest for the given problem statement and I found that Decision Tree and Random Forest models are the best suited for the given problem statement."
   ]
  }
 ],
 "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}