BCG PowerCo Churn Model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# BCG PowerCo Churn Evaluation Using Data Science  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "Created on Tue Dec 29 12:22:05 2020\n",
    "\n",
    "@author: Albert\n",
    "This file was created to show exploratory data analysis for BCG's virtuaal experience. Thank You. \n",
    "\"\"\"\n",
    "import pandas as pd\n",
    "import os\n",
    "import datetime\n",
    "import numpy as np\n",
    "import seaborn as sns #visualisation\n",
    "import matplotlib.pyplot as plt #visualisation\n",
    "from scipy import stats\n",
    "sns.set(color_codes=True)\n",
    "#os.getcwd()\n",
    "\n",
    "# EDA and Data cleaning \n",
    "#   1. Loading data through pandas\n",
    "#       A. Loading three dataframes\n",
    "#   2. Data cleaning\n",
    "#      A. Drop Duplicates - remove duplicate id's \n",
    "#      B. Remove Unwanted Columns\n",
    "#      C. Remove large n/a value columns\n",
    "#      D. Combine all Dataframes\n",
    "#   3. General statistics of a dataframe\n",
    "#       A. Data types\n",
    "#       B. Data statistics\n",
    "#   4. Format/ Convert data\n",
    "#       A. Dates\n",
    "#   5. Data visualization\n",
    "#       A. Churn percentage\n",
    "#       B. Detecting outliers using boxplots\n",
    "#       C. Using a correlation matrix\n",
    "#       D. Looking at p-values\n",
    "#       E. Final Thoughts EDA\n",
    "#   6. Saving the Cleaned Data as an Excel File\n",
    "\n",
    "\n",
    "# Feature Engineering and Predictive Modeling\n",
    "#   7. Data Evaluation\n",
    "#       A. Plotting distributions to look for skewdness\n",
    "#   8. Detecting outliers\n",
    "#       A. Standard Deviation\n",
    "#   9. Logistic Regression and Model Testing\n",
    "#       A. Creating Test and Training sets\n",
    "#       B. Evaluating the Model\n",
    "#       C. Confusion Matrix\n",
    "#       D. Churn Prediction Model Results\n",
    "#   10. Discount Hypothesis Test\n",
    "#       A. Identify Variable weights\n",
    "#       B. Provide 20% discount\n",
    "#       C. Run Model with 20% discount\n",
    "#       D. Hypothesis Results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA and Data Cleaning\n",
    "### 1. Loading data through pandas  \n",
    "    A. Loading three dataframes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                 0\n",
       "price_date         0\n",
       "price_p1_var    1359\n",
       "price_p2_var    1359\n",
       "price_p3_var    1359\n",
       "price_p1_fix    1359\n",
       "price_p2_fix    1359\n",
       "price_p3_fix    1359\n",
       "dtype: int64"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train = pd.read_csv ('/Users/yasminrassouli/Documents/BCG Data Science/ml_case_training_data.csv')\n",
    "df_hist = pd.read_csv ('/Users/yasminrassouli/Documents/BCG Data Science/ml_case_training_hist_data.csv')\n",
    "df_churn = pd.read_csv ('/Users/yasminrassouli/Documents/BCG Data Science/ml_case_training_output.csv')\n",
    "\n",
    "df_train.isnull().sum()\n",
    "df_hist.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Data Cleaning\n",
    "    A. Drop Duplicates - Remove Duplicate id's \n",
    "    B. Remove Unwanted Columns\n",
    "    C. Remove Large n/a Value Columns\n",
    "    D. Combine all Dataframes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(16043, 29)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" A. DROP DUPLICATES - remove duplicate id's \"\"\"\n",
    "df_train = df_train.drop_duplicates(subset=['id'])\n",
    "df_hist = df_hist.drop_duplicates(subset=['id'])\n",
    "\n",
    "\"\"\" point out columns with large numbers of missing values in the dataframe df_missing \"\"\"\n",
    "df_missing = df_train.columns[df_train.isnull().any()]\n",
    "\n",
    "\"\"\" B. remove 'activity_new', 'channel_sales', 'campaign_disc_ele','date_first_activ', 'forecast_base_bill_ele', \n",
    "                        'forecast_base_bill_year', 'forecast_bill_12m', 'forecast_cons', 'forecast_discount_energy', \n",
    "                        'forecast_price_energy_p1', 'forecast_price_energy_p2','forecast_price_pow_p1' columns from df \"\"\"\n",
    "\n",
    "df_train = df_train.drop(columns = ['activity_new', 'campaign_disc_ele','date_first_activ', 'forecast_base_bill_ele', \n",
    "                        'forecast_base_bill_year', 'forecast_bill_12m', 'forecast_cons', 'channel_sales', 'date_modif_prod', 'date_renewal', 'origin_up'])\n",
    "\n",
    "\"\"\" C. Remove columns with missing values are listed in the dataframe df_missing \"\"\"\n",
    "df_train = df_train.dropna(subset=['date_end', 'pow_max', 'margin_gross_pow_ele', 'margin_net_pow_ele', 'net_margin'])\n",
    "df_hist = df_hist.dropna(subset=['price_p1_var', 'price_p2_var', 'price_p3_var', 'price_p1_fix', 'price_p2_fix', 'price_p3_fix'])\n",
    "\n",
    "\"\"\" D. Combining the three dataframes train hist and churn into one df \"\"\"\n",
    "\"\"\"merge output dataframe with case_training by id to include churn values in case_training df\"\"\"\n",
    "df = pd.merge(df_train, df_hist, on='id')\n",
    "df = pd.merge(df, df_churn, on='id')\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['cons_12m']=df['cons_12m'].mask(df['cons_12m']==0).fillna(df['cons_12m'].mean())\n",
    "df['cons_last_month']=df['cons_last_month'].mask(df['cons_last_month']==0).fillna(df['cons_last_month'].mean())\n",
    "df['forecast_cons_12m']=df['forecast_cons_12m'].mask(df['forecast_cons_12m']==0).fillna(df['forecast_cons_12m'].mean())\n",
    "df['forecast_meter_rent_12m']=df['forecast_meter_rent_12m'].mask(df['forecast_meter_rent_12m']==0).fillna(df['forecast_meter_rent_12m'].mean())\n",
    "df['forecast_cons_year']=df['forecast_cons_year'].mask(df['forecast_cons_year']==0).fillna(df['forecast_cons_year'].mean())\n",
    "df['imp_cons']=df['imp_cons'].mask(df['imp_cons']==0).fillna(df['imp_cons'].mean())\n",
    "df['forecast_discount_energy']=df['forecast_discount_energy'].mask(df['forecast_discount_energy']==0).fillna(df['forecast_discount_energy'].mean())\n",
    "df['forecast_price_energy_p1']=df['forecast_price_energy_p1'].mask(df['forecast_price_energy_p1']==0).fillna(df['forecast_price_energy_p1'].mean())\n",
    "df['cons_gas_12m']=df['cons_gas_12m'].mask(df['cons_gas_12m']==0).fillna(df['cons_gas_12m'].mean())\n",
    "df['forecast_price_energy_p2']=df['forecast_price_energy_p2'].mask(df['forecast_price_energy_p2']==0).fillna(df['forecast_price_energy_p2'].mean())\n",
    "df['forecast_price_pow_p1']=df['forecast_price_pow_p1'].mask(df['forecast_price_pow_p1']==0).fillna(df['forecast_price_pow_p1'].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Series([], dtype: float64)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" ---------------------------- EXPLORATORY DATA ANALYSIS ---------------------------- \"\"\"\n",
    "\n",
    "\"\"\"CHECK IF CURRENT MISSING VALUES MAKE SENSE\"\"\"\n",
    "df_missing = df.columns[df.isnull().any()]\n",
    "df[df_missing].isna().sum() "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. General Statistics\n",
    "    A. Data Types in the Dataframe\n",
    "    B. Basic Statistics of the Dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                           object\n",
       "cons_12m                    float64\n",
       "cons_gas_12m                float64\n",
       "cons_last_month             float64\n",
       "date_activ                   object\n",
       "date_end                     object\n",
       "forecast_cons_12m           float64\n",
       "forecast_cons_year          float64\n",
       "forecast_discount_energy    float64\n",
       "forecast_meter_rent_12m     float64\n",
       "forecast_price_energy_p1    float64\n",
       "forecast_price_energy_p2    float64\n",
       "forecast_price_pow_p1       float64\n",
       "has_gas                      object\n",
       "imp_cons                    float64\n",
       "margin_gross_pow_ele        float64\n",
       "margin_net_pow_ele          float64\n",
       "nb_prod_act                   int64\n",
       "net_margin                  float64\n",
       "num_years_antig               int64\n",
       "pow_max                     float64\n",
       "price_date                   object\n",
       "price_p1_var                float64\n",
       "price_p2_var                float64\n",
       "price_p3_var                float64\n",
       "price_p1_fix                float64\n",
       "price_p2_fix                float64\n",
       "price_p3_fix                float64\n",
       "churn                         int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>net_margin</th>\n",
       "      <th>num_years_antig</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1.604300e+04</td>\n",
       "      <td>1.604300e+04</td>\n",
       "      <td>1.604300e+04</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "      <td>16043.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.950453e+05</td>\n",
       "      <td>5.804355e+04</td>\n",
       "      <td>2.574477e+04</td>\n",
       "      <td>2417.188907</td>\n",
       "      <td>2674.201095</td>\n",
       "      <td>1.937507</td>\n",
       "      <td>73.801941</td>\n",
       "      <td>0.136091</td>\n",
       "      <td>0.076736</td>\n",
       "      <td>43.799681</td>\n",
       "      <td>...</td>\n",
       "      <td>217.795809</td>\n",
       "      <td>5.030169</td>\n",
       "      <td>20.555696</td>\n",
       "      <td>0.142563</td>\n",
       "      <td>0.054829</td>\n",
       "      <td>0.030372</td>\n",
       "      <td>43.223211</td>\n",
       "      <td>10.706412</td>\n",
       "      <td>6.464706</td>\n",
       "      <td>0.099109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>6.766314e+05</td>\n",
       "      <td>1.734228e+05</td>\n",
       "      <td>8.133254e+04</td>\n",
       "      <td>4020.667470</td>\n",
       "      <td>5057.893349</td>\n",
       "      <td>4.958649</td>\n",
       "      <td>77.221345</td>\n",
       "      <td>0.025565</td>\n",
       "      <td>0.022828</td>\n",
       "      <td>3.878491</td>\n",
       "      <td>...</td>\n",
       "      <td>366.513443</td>\n",
       "      <td>1.675438</td>\n",
       "      <td>21.486067</td>\n",
       "      <td>0.022859</td>\n",
       "      <td>0.050371</td>\n",
       "      <td>0.035854</td>\n",
       "      <td>5.323104</td>\n",
       "      <td>12.835470</td>\n",
       "      <td>7.773061</td>\n",
       "      <td>0.298817</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-1.252760e+05</td>\n",
       "      <td>-3.037000e+03</td>\n",
       "      <td>-9.138600e+04</td>\n",
       "      <td>-16689.260000</td>\n",
       "      <td>-85627.000000</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>-242.960000</td>\n",
       "      <td>0.000600</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>-0.122184</td>\n",
       "      <td>...</td>\n",
       "      <td>-4148.990000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>6.081500e+03</td>\n",
       "      <td>3.192405e+04</td>\n",
       "      <td>1.436500e+03</td>\n",
       "      <td>554.560000</td>\n",
       "      <td>1156.000000</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>16.430000</td>\n",
       "      <td>0.115237</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>51.915000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>12.500000</td>\n",
       "      <td>0.127566</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.587800e+04</td>\n",
       "      <td>3.192405e+04</td>\n",
       "      <td>7.762000e+03</td>\n",
       "      <td>1256.970000</td>\n",
       "      <td>1904.733092</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>24.700000</td>\n",
       "      <td>0.142881</td>\n",
       "      <td>0.086027</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>119.670000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>13.856000</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.084991</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>5.248250e+04</td>\n",
       "      <td>3.192405e+04</td>\n",
       "      <td>1.942908e+04</td>\n",
       "      <td>2688.030000</td>\n",
       "      <td>1993.000000</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>131.480000</td>\n",
       "      <td>0.146348</td>\n",
       "      <td>0.098837</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>275.740000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>19.800000</td>\n",
       "      <td>0.152445</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.609711e+07</td>\n",
       "      <td>4.188440e+06</td>\n",
       "      <td>4.538720e+06</td>\n",
       "      <td>103801.930000</td>\n",
       "      <td>175375.000000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>2411.690000</td>\n",
       "      <td>0.273963</td>\n",
       "      <td>0.195975</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>...</td>\n",
       "      <td>24570.650000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>500.000000</td>\n",
       "      <td>0.280700</td>\n",
       "      <td>0.229298</td>\n",
       "      <td>0.114102</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>36.490692</td>\n",
       "      <td>16.791555</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "count  1.604300e+04  1.604300e+04     1.604300e+04       16043.000000   \n",
       "mean   1.950453e+05  5.804355e+04     2.574477e+04        2417.188907   \n",
       "std    6.766314e+05  1.734228e+05     8.133254e+04        4020.667470   \n",
       "min   -1.252760e+05 -3.037000e+03    -9.138600e+04      -16689.260000   \n",
       "25%    6.081500e+03  3.192405e+04     1.436500e+03         554.560000   \n",
       "50%    1.587800e+04  3.192405e+04     7.762000e+03        1256.970000   \n",
       "75%    5.248250e+04  3.192405e+04     1.942908e+04        2688.030000   \n",
       "max    1.609711e+07  4.188440e+06     4.538720e+06      103801.930000   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "count        16043.000000              16043.000000             16043.000000   \n",
       "mean          2674.201095                  1.937507                73.801941   \n",
       "std           5057.893349                  4.958649                77.221345   \n",
       "min         -85627.000000                  0.990265              -242.960000   \n",
       "25%           1156.000000                  0.990265                16.430000   \n",
       "50%           1904.733092                  0.990265                24.700000   \n",
       "75%           1993.000000                  0.990265               131.480000   \n",
       "max         175375.000000                 50.000000              2411.690000   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "count              16043.000000              16043.000000   \n",
       "mean                   0.136091                  0.076736   \n",
       "std                    0.025565                  0.022828   \n",
       "min                    0.000600                  0.052952   \n",
       "25%                    0.115237                  0.052952   \n",
       "50%                    0.142881                  0.086027   \n",
       "75%                    0.146348                  0.098837   \n",
       "max                    0.273963                  0.195975   \n",
       "\n",
       "       forecast_price_pow_p1  ...    net_margin  num_years_antig  \\\n",
       "count           16043.000000  ...  16043.000000     16043.000000   \n",
       "mean               43.799681  ...    217.795809         5.030169   \n",
       "std                 3.878491  ...    366.513443         1.675438   \n",
       "min                -0.122184  ...  -4148.990000         1.000000   \n",
       "25%                40.606701  ...     51.915000         4.000000   \n",
       "50%                44.311378  ...    119.670000         5.000000   \n",
       "75%                44.311378  ...    275.740000         6.000000   \n",
       "max                59.444710  ...  24570.650000        16.000000   \n",
       "\n",
       "            pow_max  price_p1_var  price_p2_var  price_p3_var  price_p1_fix  \\\n",
       "count  16043.000000  16043.000000  16043.000000  16043.000000  16043.000000   \n",
       "mean      20.555696      0.142563      0.054829      0.030372     43.223211   \n",
       "std       21.486067      0.022859      0.050371      0.035854      5.323104   \n",
       "min        1.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%       12.500000      0.127566      0.000000      0.000000     40.565969   \n",
       "50%       13.856000      0.148825      0.084991      0.000000     44.266931   \n",
       "75%       19.800000      0.152445      0.103395      0.071536     44.266931   \n",
       "max      500.000000      0.280700      0.229298      0.114102     59.444710   \n",
       "\n",
       "       price_p2_fix  price_p3_fix         churn  \n",
       "count  16043.000000  16043.000000  16043.000000  \n",
       "mean      10.706412      6.464706      0.099109  \n",
       "std       12.835470      7.773061      0.298817  \n",
       "min        0.000000      0.000000      0.000000  \n",
       "25%        0.000000      0.000000      0.000000  \n",
       "50%        0.000000      0.000000      0.000000  \n",
       "75%       24.339581     16.226389      0.000000  \n",
       "max       36.490692     16.791555      1.000000  \n",
       "\n",
       "[8 rows x 24 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Data Format\n",
    "    A. Data Type Conversion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\" A. DATA TYPE CONVERSION \"\"\"\n",
    "# Date columns are in object format and can be converted to integer format but since this \n",
    "# analysis is based on price and churn, we will not be doing any conversion on dates. Instead,\n",
    "# we will convert 'true', 'false' values into binary '1', '0' values respectively for all\n",
    "# applicable cases.\n",
    "\n",
    "df[\"has_gas\"]= df[\"has_gas\"].replace([\"t\", \"f\"],[1,0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5. Data Visualization\n",
    "    A. Churn Data Visualization\n",
    "    B. Detecting Outliers Using Box Plots\n",
    "    C. Using a Correlation Matrix\n",
    "    D. Looking at p-values\n",
    "    E. Final Thoughts EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\"\"\" A. CHURN DATA VISUALIZATION\"\"\"\n",
    "\n",
    "# Calculate churn percentage (it's about 10%)\n",
    "churn = df[[\"id\",\"churn\"]]\n",
    "churn_total = churn.groupby(churn[\"churn\"]).count()\n",
    "churn_percentage = churn_total/churn_total.sum()*100\n",
    "\n",
    "def plot_stacked_bars(dataframe, title_, size_=(18, 10), rot_=0, legend_=\"upper right\"):\n",
    "    \"\"\"\n",
    "     Plot stacked bars with annotations\n",
    "     \"\"\"\n",
    "    ax = dataframe.plot(kind=\"bar\",\n",
    "                         stacked=True,\n",
    "                         figsize=size_,\n",
    "                         rot=rot_,\n",
    "                         title=title_)\n",
    "     # Annotate bars\n",
    "    annotate_stacked_bars(ax, textsize=14)\n",
    "     # Rename legend\n",
    "    plt.legend([\"Retention\", \"Churn\"], loc=legend_)\n",
    "     # Labels\n",
    "    plt.ylabel(\"Company base (%)\")\n",
    "    plt.show()\n",
    "\n",
    "def annotate_stacked_bars(ax, pad=0.99, colour=\"white\", textsize=13):\n",
    "    \"\"\"\n",
    "    Add value annotations to the bars\n",
    "    \"\"\"\n",
    "    # Iterate over the plotted rectanges/bars\n",
    "    for p in ax.patches:\n",
    "        # Calculate annotation\n",
    "        value = str(round(p.get_height(),1))\n",
    "        # If value is 0 do not annotate\n",
    "        if value == '0.0':\n",
    "            continue\n",
    "        ax.annotate(value,\n",
    "                    ((p.get_x()+ p.get_width()/2)*pad-0.05, (p.get_y()+p.get_height()/2)*pad),\n",
    "                    color=colour,\n",
    "                    size=textsize,\n",
    "                    )\n",
    "\n",
    "plot_stacked_bars(churn_percentage.transpose(),\"Churning status\", (5,5), legend_=\"lower right\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='churn'>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\"\"\" B. DETECTING OUTLIERS USING BOX PLOTS\"\"\"\n",
    "#sns.boxplot(x=df['cons_gas_12m'])\n",
    "#sns.boxplot(x=df['forecast_cons_year'])\n",
    "#sns.boxplot(x=df['nb_prod_act'])\n",
    "#sns.boxplot(x=df['num_years_antig'])\n",
    "sns.boxplot(x=df['margin_gross_pow_ele'])                       \n",
    "sns.boxplot(x=df['margin_gross_pow_ele'])\n",
    "sns.boxplot(x=df['margin_net_pow_ele'])\n",
    "sns.boxplot(x=df['price_p2_var'])\n",
    "sns.boxplot(x=df['price_p3_var'])\n",
    "sns.boxplot(x=df['price_p2_fix'])\n",
    "sns.boxplot(x=df['price_p3_fix'])\n",
    "sns.boxplot(x=df['churn'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='cons_12m'>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# has significant outlier in boxplot\n",
    "sns.boxplot(x=df['cons_12m'])                             "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='net_margin'>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# has significant outlier in boxplot\n",
    "sns.boxplot(x=df['net_margin']) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='cons_last_month'>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAAEJCAYAAAA3l/RWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAQ40lEQVR4nO3dfWyN9//H8VfP6fGtMq0qnQT77o6yjUw2VLsK29yV0Zlt7KaLzU1DZH6buYkhxH3CFPmaDcuwUeo22ZYxi4a62WpzF5qauLdpmI6Joufz+0OcuZ2i3lfxfPylp+dcn3cvybPXuc7pdcKcc04AABM+rwcAgPsJ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADIXf6A5//vm3gsHSeytvlSoVdezYqVLb3t2K/XAB++EC9sO9sw98vjBVrlzhut+/YXSDQVeq0b24TbAfLmI/XMB+uD/2AacXAMAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDN/y4ntI2dOhQHTnyu558soG6dn3benkA8JR5dPPy8nT69GlFRUVbLw0AnuP0AgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgyje66ddk6f/68JKmw8ITWrcu2XB4APGca3bVr11wS3UKtXbvGcnkA8BynFwDAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMhd+JjZ448aemT58iv7+cdu7cqoYNn1GfPv932X2cc8rL26lu3bqW6tpRUZVVWPjnTT/O7w/XZ599GZo9Pb2voqKir3nf9PRuKio6I0mqUaOWRowYezsjAyhjStKBW3VHjnRXrFii/Pw87dy5VZK0efPPV93n3Lmzd2LpWwquJBUXn5f0z+zLly++7n0vBleSDh7cf0vrASi7StKBW1Xq0T1x4k+tXbtGzrnLbp86deJlXxcXF5f20rftvffeCs2+dm22CgtPXHWf9PRuV902dOhAg+kAWLi0YdfrwO0o9eiuWLFEwaC76vbNm39WYeEJBYPB0l6y1ASDxaHZg8HgNX/LXXqUexFHu8C949KGXa8Dt6PUo7t+/brQU/W70cXZi4vPa/36dR5PA8DapQ27Ex0o9egmJCTK77/263NRUdHy+cr2GyYuzu73hyshIdHjaQBYu7Rhd6IDpV7A9u1T5fOFXXV7w4bPlPZSpc7n84dm9/l8eumll6+6z3/+E3HVbTVq1LrjswGwcWnDrteB21Hq0Y2OrqykpGYKC7s8vFe+Zczv95f20rft88/nhGZPSkq+5ltF/ve/WVfdxlvGgHvHpQ27Xgduxx15rt++faoef7yO6tatL+naR7mBQLk7sbSioirf0uMuPp24OPu//Xa79GiXo1zg3lOSDtyqMHfle7uucOzYqWu+G+FWjBs3Uvn5eQoGgypfPlK1aj2kAQM+LpVt322qVn1ABQUnvR7Dc+yHC9gP984+8PnCVKVKxet/33AWALjvEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMmUY3KamZwsPDJUlRUVFKSmpmuTwAeM40uomJyZdEN1qJicmWywOA5zi9AACGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYMo9unTp19OCD1VWz5kPWSwOA58KtFxwxYoQKCk5aLwsAZQKnFwDAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQzf8uB6fL6zUF70T27wbsR8uYD9cwH64N/bBjX6GMOecM5oFAO57nF4AAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDptFdsWKF2rZtq5YtW2revHmWS5cpp06dUrt27XTw4EGvR/HM1KlTlZKSopSUFI0fP97rcTwzefJktW3bVikpKZo9e7bX43hq3LhxGjhwoNdj3HFm0f3jjz80adIkffXVV1q6dKkWLFig3bt3Wy1fZmzZskVdunTR3r17vR7FMzk5OVq7dq2WLFmipUuXaseOHVq5cqXXY5nbtGmTNmzYoOXLlysrK0tz5szRnj17vB7LE+vXr9eSJUu8HsOEWXRzcnLUpEkTRUdHKzIyUq1atdJ3331ntXyZkZmZqWHDhqlatWpej+KZqlWrauDAgSpXrpwCgYAeffRRHT582OuxzDVq1EhffvmlwsPDdezYMRUXFysyMtLrscydOHFCkyZNUq9evbwexcQNrzJWWo4ePaqqVauGvq5WrZq2bt1qtXyZMWrUKK9H8Nzjjz8e+vfevXv17bff6uuvv/ZwIu8EAgFlZGRo1qxZat26teLi4rweydzQoUPVr18/HTlyxOtRTJgd6QaDQYWF/XPJM+fcZV/j/pOfn69u3brpo48+0n//+1+vx/FM3759tX79eh05ckSZmZlej2Nq4cKFql69uhISErwexYzZke6DDz6on3/+OfR1QUHBff0U+36Xm5urvn37avDgwUpJSfF6HE/89ttvOnv2rOrWravy5curZcuWysvL83osU998840KCgrUoUMHFRYW6vTp0xo9erQGDx7s9Wh3jFl0mzZtqilTpuj48eMqX768vv/+e40cOdJqeZQhR44cUe/evTVp0qT76gjnSgcPHlRGRkbo1MoPP/ygTp06eTyVrUvfsbF48WJt2rTpng6uZBjduLg49evXT2+//bbOnTunV155RfXr17daHmXIzJkzVVRUpLFjx4Zue/3119WlSxcPp7LXrFkzbd26VR07dpTf71fLli3v26P++wmfHAEAhviLNAAwRHQBwBDRBQBDRBcADBFdALhCSS9KtWfPHr311lt66aWX9O6776qwsPCG2ya6AHCJkl6Uyjmn9PR0de/eXcuXL1fdunU1Y8aMG26f6MJzixcvVs+ePW/58UOGDNH27dtLcaKbt3DhwtDlSqdMmaIRI0Z4Og9u3bUuSrV06VKlpqaqQ4cOGjx4sIqKirRjxw5FRkYqOTlZktSrVy+98cYbN9w+0cVdLycnR16/3Tw3N1dnzpzxdAaUjlGjRumZZ54JfZ2fn6/MzEzNnz9fy5YtU5UqVTRz5kzt379fsbGxGjx4sFJTUzVs2LASXSXO7C/SUPYtWrRIs2fPls/nU+XKlTVu3DhlZ2drzpw58vl8io2N1ccff6yHH35YAwcOVMWKFZWXl6fff/9dderU0bhx41ShQgVlZGRo5cqVCgQCqly5ssaMGVPi62z8+uuvmjBhgs6ePauCggI1bdpUo0eP1vnz5zVy5Eht3rxZgUBANWrU0JgxYzRjxgwdPXpUH374ocaPH68GDRpcc7sHDx5UWlqaEhMTtX37dhUXF6tv375asGCB9uzZoyeffFITJ06Uz+fTqlWrNHXqVAWDQVWoUEGDBg1S/fr1NWXKFB06dEgFBQU6dOiQ4uLiNGHCBG3ZskWrV6/WunXrFBERIemfc30FBQWKjY3VxIkTudbIXWrjxo3at2+fXn31VUnSuXPnVK9ePdWoUUObNm3S3Llz9dRTT+mTTz7R2LFjL/tLy2tygHNu586drnHjxu7w4cPOOedmz57tWrZs6V544QV37Ngx55xzWVlZrk2bNi4YDLoBAwa41157zRUVFbmzZ8+6jh07ukWLFrnDhw+7hg0buqKiIuecczNnznQrV67817WzsrJcjx49nHPO9evXz23YsME559ypU6dc48aN3bZt29xPP/3kWrdu7YLBoHPOufHjx7vc3FznnHPNmzd3W7du/dc1Dhw44GrXru1WrVrlnHNu6NChrnnz5u7kyZPuzJkzLjEx0eXm5rrdu3e7pk2buv379zvnnMvJyXGJiYnu5MmTLiMjwz3//PPu5MmTzjnnevbs6SZPnuycc27AgAHu888/d845l5GR4Vq0aBHab+np6W7q1Kkl+n9A2dG8eXN34MAB98UXX7iRI0eGbj916pQrLCx0OTk5rn379qHb8/PzXZs2bW64XY50IenClfuTkpJUvXp1SdI777yjo0ePKhAIKCYmRpL08ssva9SoUaFXdJ977jmVK1dOklS7dm0VFhYqLi5O8fHxSk1NVXJyspKTk2/qojZjx45Vdna2pk+frj179qioqEinT59WfHy8/H6/OnfurKSkJLVq1eqmr90RCATUokULSVKtWrX09NNPq2LFipIuXN+5sLBQO3fuVJMmTVSzZk1JUkJCgmJiYkLnjBs1ahR6TL169a77anViYmJov8XHx+v48eM3NSvKjsaNG2vWrFlKT09XTEyMhg8frlq1aql79+46fvy4du3apfj4eK1evVpPPPHEDbfHOV1Ikvx+/2XXNz5z5owOHDhw1f2cczp//rwkhZ5KS1JYWJicc/L5fJo7d67GjBmj6OhojR49+qY+A+3NN9/UmjVr9Mgjj6h3796qVq2anHOqVKmSli1bpgEDBsjv9+v999+/6c/ZCwQCl/2MgUDgqvtced3nkvzM1xIeHl6i+6Hsi4+PV58+fZSWlqaUlBQFg0H16NFDERERmjZtmoYMGaKUlBRt3LixRJ/xxpEuJF34bX7x/Gi1atU0f/58rVmzRrt27VJaWppiYmKUlZWl6OhoPfTQQ9fdzq5du/TBBx8oMzNT9evXV2xsrJYuXVqiGf766y9t27ZNn332maKiorRx40bt379fwWBQP/74o2bNmqXZs2fr2WeflXMudPTp9/tDUbxdCQkJmjZtmg4cOKCaNWuGLi7eoEED/fLLL9d9XGnOgLJh9erVoX937txZnTt3vuo+DRo00KJFi25qu0QXkqQ6deqof//+eu+99yRd+ByzlStXatWqVUpLS1MwGFRMTIw+/fRT+XzXf4IUHx+vNm3aqFOnToqMjFRERISGDBlSohkqVaqkHj16KDU1VZGRkYqLi1PDhg21b98+de7cWdnZ2WrXrp0iIyMVFRUVuh7ziy++qP79+2v48OFKSkq6rf3w2GOPadiwYerTp4+Ki4sVERGh6dOn64EHHvjXxyUnJ9/4BRRAXNoRAExxpAsTXbt21d9//33N782bNy/04tTtGD16tDZu3HjN7w0aNEhNmjS57TWA28WRLgAY4t0LAGCI6AKAIaILAIaILgAYIroAYOj/AQHJWFnjiUa4AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# has significant outlier in boxplot\n",
    "sns.boxplot(x=df['cons_last_month'])                          "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='price_p1_var'>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# has significant outlier in boxplot\n",
    "sns.boxplot(x=df['price_p1_var'])                          "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='price_p1_fix'>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAAEJCAYAAAA3l/RWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAP8UlEQVR4nO3df0yVdf/H8Rc//dHXJG41sjmXzWXLZa6mIUylr0PjQPhrpZi6USvvWTrbnXA7v7mWGDKX+aOmy7JZWWppCkuzTO8CkbKVZklyawoqlZkHshDw8Pn+cU+6NQuVi/fh4PPxVwcO1+fzjrMnF1enizDnnBMAwER4sDcAAFcTogsAhoguABgiugBgiOgCgCGiCwCGiC4AGIps6gmnTv2qhobLfyvv3/72Pzp58vQVbao1akvzMEvrxCyt1+XMEx4epuuuu+ZPP99kdBsa3BVF99zXtiVtaR5maZ2YpfXyah4uLwCAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhpr8cz0AbK1evUoVFUc8OVZVlV+xsV30j3/805PjofmILtDKVFQc0bdl/1ZE+5hmHyvw20+qqqpq/qbgGaILtEIR7WPUsef/Nvs4v3z7jge7gZe4pgsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiCwCGiC4AGCK6AGCI6AKAIaILAIaILgAYIroAYIjoAoAhogsAhoguABgiugBgiOgCgCGiC7SQoqKPVVT0cbC30aRQ2WdbERnsDQBtVWHhvyRJCQmDg7yTvxYq+2wrONMFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMBTZEgfNzMxo/OdXXlndEku0OX7/KS1btkR///s0de4cE+ztAEFVULBR69ev0f33j9eIEWkma/53t855+uln1aNHT0/X4Uy3lcjP36Cysm+1adP6YG8FCLr169dIktaufTOo+1i+fKnnx/Q8uhf+tLjYTw+cz+8/pcLCf8k5p8LCj1VV5Q/2loCgKSjYeN7jLVvyW3zNP+vU8ePHVFFxxNO1WuTyAi5Pfv4GNTQ4SVJDQ4M2bVqviRMzg7wrNFdVlV9VVVWaP/+ZS3p+VFSE6usDKi8/ooZAhDebcA2qrT3zl3soLz+izp07e7OeB86d5Z6zdu2bZpcYLmb58qVavnyZZ8fj8kIrUFxcpEDgrCQpEDir4uKiIO8IwDnHjx/z9Hic6bYC8fEJ+vjjHQoEzioiIlLx8QnB3hI80LlzjDp3jlFW1v9d0vO7du2kEyd+0fz5z+jfFT95s4mwcLVrF/WXe7jUM/GrVffuN3p6PM50W4G0tFEKDw+TJIWHh+u++0YHeUdA8Iwe/cB5j++/f3yQdvIfjz76mKfH8zy6F75FjLeMNS0m5jolJg5RWFiYEhMH85YxXNVSU9PPe2xxPffPOtW9+428ZaytSksbpd69b+EsF9DvZ7tt7SxXaqFruq+8srrx+hQuTUzMdcrOfirY2wBahdTU9D+c8bY0q9/KOdMFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAEORwd4A0FYlJg4J9hYuSajss60gukALSUgYHOwtXJJQ2WdbweUFADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMBQZ7A0A+KPAGb9+O7Kt+QdqOCspqvnHgWeILtDK9OjR07NjVVVFKja2i2fHQ/MRXaCVyciYFOwtoAVxTRcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQ0QXAAwRXQAwRHQBwBDRBQBDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMNTkn+sJDw+74oM352tbo7Y0D7O0TszSel3qPE09L8w557zYEACgaVxeAABDRBcADBFdADBEdAHAENEFAENEFwAMEV0AMER0AcAQ0QUAQy0S3fz8fKWkpCg5OVlvvPFGSyzR4k6fPq3U1FQdPXpUkrRz506lpaUpOTlZCxcuDPLuLt3SpUvl8/nk8/mUl5cnKXRnWbRokVJSUuTz+bRy5UpJoTvLOfPnz1d2drak0J5l4sSJ8vl8Sk9PV3p6uvbs2ROy83z00UcaPXq07r33Xs2dO1eSx98b57Hvv//eJSUluVOnTrlff/3VpaWlubKyMq+XaVFffvmlS01NdbfddpurqKhwNTU1bsiQIa68vNzV19e7zMxMt2PHjmBvs0lFRUXugQcecLW1ta6urs5NmjTJ5efnh+QsJSUlbty4ca6+vt7V1NS4pKQkt3///pCc5ZydO3e6gQMHuqysrJB9jTnnXENDg0tMTHT19fWNHwvVecrLy11iYqKrrKx0dXV1bvz48W7Hjh2ezuL5me7OnTt19913KyYmRh07dtTw4cO1ZcsWr5dpUWvXrtWcOXPUrVs3SdLevXvVs2dP9ejRQ5GRkUpLSwuJmbp27ars7GxFR0crKipKN998sw4fPhySswwYMECrVq1SZGSkTp48qUAgoOrq6pCcRZL8fr8WLlyoKVOmSArd15gkHTp0SJKUmZmp++67T6+//nrIzvPBBx8oJSVFcXFxioqK0sKFC9WhQwdPZ/E8uj/++KO6du3a+Lhbt2764YcfvF6mReXk5Oiuu+5qfByqM/Xu3Vt33HGHJOnw4cPavHmzwsLCQnIWSYqKitLixYvl8/kUHx8fst8XSXrqqac0Y8YMXXvttZJC9zUmSdXV1YqPj9cLL7ygV199VW+99ZaOHz8ekvMcOXJEgUBAU6ZMUXp6ulavXu3598bz6DY0NCgs7PdbmznnznscikJ9prKyMmVmZmrmzJnq0aNHSM8ybdo0FRcXq7KyUocPHw7JWdatW6cbbrhB8fHxjR8L5ddY//79lZeXp06dOik2NlZjx47V4sWLQ3KeQCCg4uJizZs3T2vWrNHevXtVUVHh6SxN3k/3csXFxWn37t2Nj0+cONH4a3qoiouL04kTJxofh9JMn3/+uaZNm6ZZs2bJ5/Pp008/DclZDh48qLq6Ot16663q0KGDkpOTtWXLFkVERDQ+J1Rmee+993TixAmlp6erqqpKv/32m44dOxaSs0jS7t27VV9f3/hDxDmnG2+8MSRfZ126dFF8fLxiY2MlScOGDfP8deb5me6gQYNUXFysn3/+WTU1Ndq6dasGDx7s9TKm+vXrp++++67xV4+CgoKQmKmyslJTp07VggUL5PP5JIXuLEePHtXs2bNVV1enuro6bdu2TePGjQvJWVauXKmCggJt3LhR06ZN0z333KMVK1aE5CyS9MsvvygvL0+1tbU6ffq0NmzYoCeeeCIk50lKSlJhYaGqq6sVCAT0ySefaMSIEZ7O4vmZ7vXXX68ZM2Zo0qRJqq+v19ixY3X77bd7vYypdu3aKTc3V48//rhqa2s1ZMgQjRgxItjbatLLL7+s2tpa5ebmNn5s3LhxITnLkCFDtHfvXo0cOVIRERFKTk6Wz+dTbGxsyM1yMaH6GpP+E6o9e/Zo5MiRamhoUEZGhvr37x+S8/Tr108PP/ywMjIyVF9fr4SEBI0fP169evXybBb+cgQAGOL/SAMAQ0QXAAwRXQAwRHQBwBDRBQBDRBctbtGiRXr33XeDsvbbb7/deH+Dpuzfv1/Dhg3T6NGjtWrVqsY7TAFe8vx9usCFpk+fbr6m3+/Xc889p/z8fA0YMOCSvmbbtm0aOHCgcnJyWnh3uJoRXVyxkpISLViwQN27d9ehQ4fUvn175ebm6qWXXpLf71dFRYWGDh2qkydPqnfv3nrooYe0Z88ezZ07VzU1NYqKitLMmTMVHx+vgwcPKicnR36/X4FAQBMnTtTYsWP/cv3s7Gy1a9dOpaWlOnnypBISEjR79mxFRUVp8+bN6tatm7KysrR9+/YmZ9m0aZPefPNNBQIBnTlzRgkJCXr//fe1aNEijRkzRhkZGZowYYLWrVunVatWae3aterQoYNX/ypxNWn2DShx1dq1a5fr06eP++yzz5xzzq1evdqNGjXKZWVlucmTJzc+Lysry61YscLV1dW5hIQEt337duecc1999ZVLTU11tbW1LiUlxe3bt88551x1dbW799573RdffPGX62dlZbmRI0e606dPu9raWjdhwgT32muvnfecd955xz3yyCOXNM/ixYvd008//YevKy0tdQMGDHA7duxwgwYNcgcPHryk4wEXwzVdNEufPn0ab4M5ZswY7d+/X36/X3feeecfnnvgwAGFh4dr6NChkqS+ffsqPz9f5eXlKi8v16xZs5Senq4HH3xQZ86c0TfffNPk+qNGjdI111yj6Ohopaenq7Cw0NP5JOmWW27RY489pkcffVRPPvmkevXq5fkauHpweQHN8t93XzonPDxcHTt2vOhzL7wl3oEDB+ScU6dOnbRx48bGj//000/q1KnTZa3vnFN4eMucR5SVlalLly6N9xgArhRnumiW0tJSlZaWSpLWrFmj/v37N96Y+0K9evVSWFiYioqKJElff/21Jk+erJtuuknt27dvjG5lZaVSU1O1b9++JtffvHmz6urqVFtbqw0bNigpKcmjyX63detWlZSUaNOmTSoqKtKHH37o+Rq4enCmi2bp0qWLnn/+eR07dkyxsbHKy8vT0qVLL/rc6OhoLVmyRPPmzVNeXp6ioqK0ZMkSRUdH68UXX1ROTo5WrFihs2fPavr06Re9RHGh9u3bKyMjQ9XV1Ro+fLjGjBnj6XyVlZWaM2eOli1bptjYWOXm5mrq1Knq27ev4uLiPF0LVwfuMoYrVlJSomeeeUYFBQVBWT87O7vxXRFAqOBMF63WoUOHNGPGjIt+7twlicuxa9cuPfvssxf93MCBAzVr1qzL3iNwuTjTBQBD/Ic0ADBEdAHAENEFAENEFwAMEV0AMER0AcDQ/wObtN5RSgbkGAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# has significant outlier in boxplot\n",
    "sns.boxplot(x=df['price_p1_fix'])                            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\"\"\" C. USING A CORRELATION MATRIX\"\"\"\n",
    "# Creating a correlation heatmap using cons_12m, consumption_last_month, and churn as rows and columns\n",
    "df[['cons_12m','cons_last_month','churn']].corr()\n",
    "sns.heatmap(df[['cons_12m','cons_last_month','churn']].corr(), annot=True, cmap = 'Reds')\n",
    "plt.show()\n",
    "\n",
    "# Creating a correlation heatmap using net_margin, margin_gross_pow_ele, margin_net_pow_ele, and churn as rows and columns\n",
    "df[['net_margin','margin_gross_pow_ele', 'margin_net_pow_ele', 'churn']].corr()\n",
    "sns.heatmap(df[['net_margin','margin_gross_pow_ele','margin_net_pow_ele', 'churn']].corr(), annot=True, cmap = 'Reds')\n",
    "plt.show()\n",
    "\n",
    "# Creating a correlation heatmap using price_p1_var, price_p2_var, price_p3_var, and churn as rows and columns\n",
    "df[['price_p1_var','price_p2_var', 'price_p3_var', 'churn']].corr()\n",
    "sns.heatmap(df[['price_p1_var','price_p2_var','price_p3_var', 'churn']].corr(), annot=True, cmap = 'Reds')\n",
    "plt.show()\n",
    "\n",
    "# Creating a correlation heatmap using price_p1_fix, price_p1_fix, price_p1_fix, and churn as rows and columns\n",
    "df[['price_p1_fix','price_p2_fix', 'price_p3_fix', 'churn']].corr()\n",
    "sns.heatmap(df[['price_p1_fix','price_p2_fix','price_p3_fix', 'churn']].corr(), annot=True, cmap = 'Reds')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PointbiserialrResult(correlation=-0.05133529999196243, pvalue=7.719793665737546e-11)\n",
      "PointbiserialrResult(correlation=-0.04695243382778795, pvalue=2.683744444977101e-09)\n",
      "PointbiserialrResult(correlation=0.02993748607669099, pvalue=0.00014915904475766123)\n",
      "PointbiserialrResult(correlation=0.08116421434888142, pvalue=7.305129917041309e-25)\n",
      "PointbiserialrResult(correlation=0.06390586489898303, pvalue=5.404969443222569e-16)\n",
      "PointbiserialrResult(correlation=0.0028309095781572768, pvalue=0.719940348067061)\n"
     ]
    }
   ],
   "source": [
    "\"\"\" D. LOOKING AT P-VALUES\"\"\"\n",
    "\n",
    "# we get low correlation for margin and consumption variables with respect to churn. Try using a point-biserial correlation to double check \n",
    "# and get a look at p values.\n",
    "\n",
    "print(stats.pointbiserialr(df['cons_12m'], df['churn']))\n",
    "print(stats.pointbiserialr(df['cons_last_month'], df['churn']))\n",
    "print(stats.pointbiserialr(df['net_margin'], df['churn']))\n",
    "print(stats.pointbiserialr(df['margin_gross_pow_ele'], df['churn']))\n",
    "print(stats.pointbiserialr(df['margin_net_pow_ele'], df['churn']))\n",
    "print(stats.pointbiserialr(df['imp_cons'], df['churn']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PointbiserialrResult(correlation=0.0006399991377887661, pvalue=0.9353969400220042)\n",
      "PointbiserialrResult(correlation=0.024356882154937202, pvalue=0.002033582261960833)\n",
      "PointbiserialrResult(correlation=0.03912184198132497, pvalue=7.16925915356387e-07)\n",
      "PointbiserialrResult(correlation=0.007230515072909119, pvalue=0.3597903613000602)\n",
      "PointbiserialrResult(correlation=0.035226599727900196, pvalue=8.086818882037972e-06)\n",
      "PointbiserialrResult(correlation=0.04025844773288037, pvalue=3.3814167375680013e-07)\n"
     ]
    }
   ],
   "source": [
    "print(stats.pointbiserialr(df['price_p1_var'], df['churn']))\n",
    "print(stats.pointbiserialr(df['price_p2_var'], df['churn']))\n",
    "print(stats.pointbiserialr(df['price_p3_var'], df['churn']))\n",
    "print(stats.pointbiserialr(df['price_p1_fix'], df['churn']))\n",
    "print(stats.pointbiserialr(df['price_p2_fix'], df['churn']))\n",
    "print(stats.pointbiserialr(df['price_p3_fix'], df['churn']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### E. Final Thoughts EDA \n",
    "Based on p-values from point-biserial correlations above, the variables cons_12m, cons_last_month, net_margin, margin_gross_pow_ele, margin_net_pow_ele,price_p2_var, price_p3_var, price_p2_fix, and price_p3_fix are statistically significant! \n",
    "\n",
    "Also, our correlation matrix is showing that there is a higher correlation between price and churn in periods 2 and 3 than any other variable. I would like to further explore the effects of pricing in periods one and two on customer churn."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6. Saving the Cleaned Data as an Excel File"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "' ---------------------------- Save final dataframe df to file directory for submission ---------------------------- '"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" ---------------------------- Save final dataframe df to file directory for submission ---------------------------- \"\"\"\n",
    "#df.to_excel('cleaned_data.xlsx')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Evaluation\n",
    "    A. Plotting Distributions to Look for Skewdness"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/yasminrassouli/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/Users/yasminrassouli/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/Users/yasminrassouli/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/Users/yasminrassouli/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/Users/yasminrassouli/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/Users/yasminrassouli/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/Users/yasminrassouli/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x3600 with 7 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_plot = df\n",
    "# Remove negative values\n",
    "df_plot.loc[df_plot.cons_12m < 0,\"cons_12m\"] = np.nan\n",
    "df_plot.loc[df_plot.cons_gas_12m < 0,\"cons_gas_12m\"] = np.nan\n",
    "df_plot.loc[df_plot.cons_last_month < 0,\"cons_last_month\"] = np.nan\n",
    "df_plot.loc[df_plot.forecast_cons_12m < 0,\"forecast_cons_12m\"] = np.nan\n",
    "df_plot.loc[df_plot.forecast_cons_year < 0,\"forecast_cons_year\"] = np.nan\n",
    "df_plot.loc[df_plot.forecast_meter_rent_12m < 0,\"forecast_meter_rent_12m\"] = np.nan\n",
    "df_plot.loc[df_plot.imp_cons < 0,\"imp_cons\"] = np.nan\n",
    "# Apply log10 transformation\n",
    "df_plot[\"cons_12m\"] = np.log10(df_plot[\"cons_12m\"]+1)\n",
    "df_plot[\"cons_gas_12m\"] = np.log10(df_plot[\"cons_gas_12m\"]+1)\n",
    "df_plot[\"cons_last_month\"] = np.log10(df_plot[\"cons_last_month\"]+1)\n",
    "df_plot[\"forecast_cons_12m\"] = np.log10(df_plot[\"forecast_cons_12m\"]+1)\n",
    "df_plot[\"forecast_cons_year\"] = np.log10(df_plot[\"forecast_cons_year\"]+1)\n",
    "df_plot[\"forecast_meter_rent_12m\"] = np.log10(df_plot[\"forecast_meter_rent_12m\"]+1)\n",
    "df_plot[\"imp_cons\"] = np.log10(df_plot[\"imp_cons\"]+1)\n",
    "\n",
    "#Plot distributions\n",
    "fig, axs = plt.subplots(nrows=7, figsize=(18,50))\n",
    "\n",
    "# Plot histograms\n",
    "sns.distplot((df_plot[\"cons_12m\"].dropna()), ax=axs[0])\n",
    "sns.distplot((df_plot[df_plot[\"has_gas\"]==1][\"cons_gas_12m\"].dropna()), ax=axs[1])\n",
    "sns.distplot((df_plot[\"cons_last_month\"].dropna()), ax=axs[2])\n",
    "sns.distplot((df_plot[\"forecast_cons_12m\"].dropna()), ax=axs[3])\n",
    "sns.distplot((df_plot[\"forecast_cons_year\"].dropna()), ax=axs[4])\n",
    "sns.distplot((df_plot[\"forecast_meter_rent_12m\"].dropna()), ax=axs[5])\n",
    "sns.distplot((df_plot[\"imp_cons\"].dropna()), ax=axs[6])\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 8. Detecting Outliers\n",
    "    A. Using Standard Deviation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>index</th>\n",
       "      <th>id</th>\n",
       "      <th>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>date_activ</th>\n",
       "      <th>date_end</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>...</th>\n",
       "      <th>num_years_antig</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_date</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>24011ae4ebbe3035111d65fa7c15bc57</td>\n",
       "      <td>5.286767</td>\n",
       "      <td>4.739944</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2013-06-15</td>\n",
       "      <td>2016-06-15</td>\n",
       "      <td>3.374651</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>43.648</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.125976</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>d29c2c54acc38ff3c0614d0a653813dd</td>\n",
       "      <td>3.668479</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2009-08-21</td>\n",
       "      <td>2016-08-30</td>\n",
       "      <td>2.280920</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>13.800</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.151367</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>764c75f661154dac3a6c254cd082ea7d</td>\n",
       "      <td>2.736397</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2010-04-16</td>\n",
       "      <td>2016-04-16</td>\n",
       "      <td>1.689841</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>13.856</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.172468</td>\n",
       "      <td>0.087881</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>bba03439a292a1e166f80264c16191cb</td>\n",
       "      <td>3.200029</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2010-03-30</td>\n",
       "      <td>2016-03-30</td>\n",
       "      <td>2.382089</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>13.200</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.153133</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>568bb38a1afd7c0fc49c77b3789b59a3</td>\n",
       "      <td>5.083990</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>2010-04-08</td>\n",
       "      <td>2016-04-08</td>\n",
       "      <td>4.036071</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>75.000</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.123086</td>\n",
       "      <td>0.100505</td>\n",
       "      <td>0.068646</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0</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>15651</th>\n",
       "      <td>16038</td>\n",
       "      <td>18463073fb097fc0ac5d3e040f356987</td>\n",
       "      <td>4.508812</td>\n",
       "      <td>4.680707</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2012-05-24</td>\n",
       "      <td>2016-05-08</td>\n",
       "      <td>3.667360</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>15.000</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15652</th>\n",
       "      <td>16039</td>\n",
       "      <td>d0a6f71671571ed83b2645d23af6de00</td>\n",
       "      <td>3.858778</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>2.260071</td>\n",
       "      <td>2012-08-27</td>\n",
       "      <td>2016-08-27</td>\n",
       "      <td>2.801191</td>\n",
       "      <td>2.260071</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>6.000</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.109747</td>\n",
       "      <td>0.096908</td>\n",
       "      <td>0.070730</td>\n",
       "      <td>58.936774</td>\n",
       "      <td>36.344726</td>\n",
       "      <td>8.334260</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15653</th>\n",
       "      <td>16040</td>\n",
       "      <td>10e6828ddd62cbcf687cb74928c4c2d2</td>\n",
       "      <td>3.265996</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>2.255273</td>\n",
       "      <td>2012-02-08</td>\n",
       "      <td>2016-02-07</td>\n",
       "      <td>2.281919</td>\n",
       "      <td>2.255273</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>15.935</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.126098</td>\n",
       "      <td>0.103975</td>\n",
       "      <td>0.070232</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15654</th>\n",
       "      <td>16041</td>\n",
       "      <td>1cf20fd6206d7678d5bcafd28c53b4db</td>\n",
       "      <td>2.120574</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2012-08-30</td>\n",
       "      <td>2016-08-30</td>\n",
       "      <td>1.308351</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>11.000</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.151367</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15655</th>\n",
       "      <td>16042</td>\n",
       "      <td>563dde550fd624d7352f3de77c0cdfcd</td>\n",
       "      <td>3.941064</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2009-12-18</td>\n",
       "      <td>2016-12-17</td>\n",
       "      <td>2.882758</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>10.392</td>\n",
       "      <td>2015-01-01</td>\n",
       "      <td>0.170590</td>\n",
       "      <td>0.087658</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>15656 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       index                                id  cons_12m  cons_gas_12m  \\\n",
       "0          1  24011ae4ebbe3035111d65fa7c15bc57  5.286767      4.739944   \n",
       "1          2  d29c2c54acc38ff3c0614d0a653813dd  3.668479      4.504132   \n",
       "2          3  764c75f661154dac3a6c254cd082ea7d  2.736397      4.504132   \n",
       "3          4  bba03439a292a1e166f80264c16191cb  3.200029      4.504132   \n",
       "4          5  568bb38a1afd7c0fc49c77b3789b59a3  5.083990      4.504132   \n",
       "...      ...                               ...       ...           ...   \n",
       "15651  16038  18463073fb097fc0ac5d3e040f356987  4.508812      4.680707   \n",
       "15652  16039  d0a6f71671571ed83b2645d23af6de00  3.858778      4.504132   \n",
       "15653  16040  10e6828ddd62cbcf687cb74928c4c2d2  3.265996      4.504132   \n",
       "15654  16041  1cf20fd6206d7678d5bcafd28c53b4db  2.120574      4.504132   \n",
       "15655  16042  563dde550fd624d7352f3de77c0cdfcd  3.941064      4.504132   \n",
       "\n",
       "       cons_last_month  date_activ    date_end  forecast_cons_12m  \\\n",
       "0             4.288474  2013-06-15  2016-06-15           3.374651   \n",
       "1             4.288474  2009-08-21  2016-08-30           2.280920   \n",
       "2             4.288474  2010-04-16  2016-04-16           1.689841   \n",
       "3             4.288474  2010-03-30  2016-03-30           2.382089   \n",
       "4             4.093457  2010-04-08  2016-04-08           4.036071   \n",
       "...                ...         ...         ...                ...   \n",
       "15651         4.288474  2012-05-24  2016-05-08           3.667360   \n",
       "15652         2.260071  2012-08-27  2016-08-27           2.801191   \n",
       "15653         2.255273  2012-02-08  2016-02-07           2.281919   \n",
       "15654         4.288474  2012-08-30  2016-08-30           1.308351   \n",
       "15655         4.288474  2009-12-18  2016-12-17           2.882758   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  ...  num_years_antig  \\\n",
       "0                3.280062                  0.990265  ...                3   \n",
       "1                3.280062                  0.990265  ...                6   \n",
       "2                3.280062                  0.990265  ...                6   \n",
       "3                3.280062                  0.990265  ...                6   \n",
       "4                4.093457                  0.990265  ...                6   \n",
       "...                   ...                       ...  ...              ...   \n",
       "15651            3.280062                  0.990265  ...                4   \n",
       "15652            2.260071                  0.990265  ...                3   \n",
       "15653            2.255273                  0.990265  ...                4   \n",
       "15654            3.280062                  0.990265  ...                3   \n",
       "15655            3.280062                  0.990265  ...                6   \n",
       "\n",
       "       pow_max  price_date  price_p1_var  price_p2_var  price_p3_var  \\\n",
       "0       43.648  2015-01-01      0.125976      0.103395      0.071536   \n",
       "1       13.800  2015-01-01      0.151367      0.000000      0.000000   \n",
       "2       13.856  2015-01-01      0.172468      0.087881      0.000000   \n",
       "3       13.200  2015-01-01      0.153133      0.000000      0.000000   \n",
       "4       75.000  2015-01-01      0.123086      0.100505      0.068646   \n",
       "...        ...         ...           ...           ...           ...   \n",
       "15651   15.000  2015-01-01      0.148825      0.000000      0.000000   \n",
       "15652    6.000  2015-01-01      0.109747      0.096908      0.070730   \n",
       "15653   15.935  2015-01-01      0.126098      0.103975      0.070232   \n",
       "15654   11.000  2015-01-01      0.151367      0.000000      0.000000   \n",
       "15655   10.392  2015-01-01      0.170590      0.087658      0.000000   \n",
       "\n",
       "       price_p1_fix  price_p2_fix  price_p3_fix  churn  \n",
       "0         40.565969     24.339581     16.226389      1  \n",
       "1         44.266931      0.000000      0.000000      0  \n",
       "2         44.266931      0.000000      0.000000      0  \n",
       "3         44.266931      0.000000      0.000000      0  \n",
       "4         40.565969     24.339581     16.226389      0  \n",
       "...             ...           ...           ...    ...  \n",
       "15651     44.266931      0.000000      0.000000      0  \n",
       "15652     58.936774     36.344726      8.334260      1  \n",
       "15653     40.565969     24.339581     16.226389      1  \n",
       "15654     44.266931      0.000000      0.000000      0  \n",
       "15655     44.266931      0.000000      0.000000      0  \n",
       "\n",
       "[15656 rows x 30 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" A. Using Standard Deviation \"\"\"\n",
    "#Dropping the outlier rows using standard deviation (note that price_p1_var and price_p1_fix were ommited)\n",
    "factor = 3\n",
    "upper_lim = df['cons_12m'].mean () + df['cons_12m'].std () * factor\n",
    "lower_lim = df['cons_12m'].mean () - df['cons_12m'].std () * factor\n",
    "\n",
    "df = df[(df['cons_12m'] < upper_lim) & (df['cons_12m'] > lower_lim)]\n",
    "\n",
    "#Dropping the outlier rows using standard deviation\n",
    "factor = 3\n",
    "upper_lim = df['net_margin'].mean () + df['net_margin'].std () * factor\n",
    "lower_lim = df['net_margin'].mean () - df['net_margin'].std () * factor\n",
    "\n",
    "df = df[(df['net_margin'] < upper_lim) & (df['net_margin'] > lower_lim)]\n",
    "\n",
    "#Dropping the outlier rows using standard deviation\n",
    "factor = 3\n",
    "upper_lim = df['cons_last_month'].mean () + df['cons_last_month'].std () * factor\n",
    "lower_lim = df['cons_last_month'].mean () - df['cons_last_month'].std () * factor\n",
    "\n",
    "df = df[(df['cons_last_month'] < upper_lim) & (df['cons_last_month'] > lower_lim)]\n",
    "df.reset_index()\n",
    "\n",
    "\n",
    "# Using this method we can see a reduction in total rows from 16042 to 15656 (96% of data remains) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 9. Logistic Regression and Model Testing\n",
    "    A. Creating Test and Training sets\n",
    "    B. Evaluating the Model\n",
    "    C. Confusion Matrix\n",
    "    D. Churn Prediction Model Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\" A. Creating Test and Training sets \"\"\"\n",
    "# Remove n/a values from columns \n",
    "\n",
    "df = df[df['cons_last_month'].notna()]\n",
    "df = df[df['cons_gas_12m'].notna()]\n",
    "df = df[df['imp_cons'].notna()]\n",
    "df = df[df['forecast_meter_rent_12m'].notna()]\n",
    "df = df[df['forecast_cons_12m'].notna()]\n",
    "df = df[df['forecast_cons_year'].notna()]\n",
    "df = df[df['forecast_meter_rent_12m'].notna()]\n",
    "df[\"churn\"] = df[\"churn\"].astype(int)\n",
    "\n",
    "# store later needed columns and drop columns deemed unnecessary for Logistic Regression Calculations\n",
    "id_number = df['id']\n",
    "df.drop(columns=['id', \"date_activ\", \"date_end\", 'price_date'],inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "Y = df[\"churn\"].values\n",
    "X = df.drop(labels = [\"churn\"],axis = 1)\n",
    "\n",
    "\"\"\" Churn Prediction Model Results\"\"\"\n",
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.20, random_state=35)\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "model = LogisticRegression(max_iter = 2000)\n",
    "result = model.fit(X_train, y_train)\n",
    "\n",
    "from sklearn import metrics\n",
    "prediction_test = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluating the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Churn predictive model accuracy is  89.82725527831093\n"
     ]
    }
   ],
   "source": [
    "# Print the prediction accuracy\n",
    "print('Churn predictive model accuracy is ', 100 * metrics.accuracy_score(y_test, prediction_test ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>num_years_antig</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>15628.00000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4.325261</td>\n",
       "      <td>4.487287</td>\n",
       "      <td>3.763063</td>\n",
       "      <td>3.074854</td>\n",
       "      <td>3.196937</td>\n",
       "      <td>1.912667</td>\n",
       "      <td>1.636982</td>\n",
       "      <td>0.136264</td>\n",
       "      <td>0.076520</td>\n",
       "      <td>43.767678</td>\n",
       "      <td>...</td>\n",
       "      <td>5.02310</td>\n",
       "      <td>19.708523</td>\n",
       "      <td>0.142776</td>\n",
       "      <td>0.054331</td>\n",
       "      <td>0.029988</td>\n",
       "      <td>43.198620</td>\n",
       "      <td>10.538924</td>\n",
       "      <td>6.396068</td>\n",
       "      <td>0.097453</td>\n",
       "      <td>0.096452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.803837</td>\n",
       "      <td>0.411195</td>\n",
       "      <td>0.777973</td>\n",
       "      <td>0.518909</td>\n",
       "      <td>0.440543</td>\n",
       "      <td>4.904279</td>\n",
       "      <td>0.467608</td>\n",
       "      <td>0.025378</td>\n",
       "      <td>0.022818</td>\n",
       "      <td>3.787615</td>\n",
       "      <td>...</td>\n",
       "      <td>1.66229</td>\n",
       "      <td>18.361513</td>\n",
       "      <td>0.022686</td>\n",
       "      <td>0.050397</td>\n",
       "      <td>0.035791</td>\n",
       "      <td>5.236536</td>\n",
       "      <td>12.768468</td>\n",
       "      <td>7.769176</td>\n",
       "      <td>0.296584</td>\n",
       "      <td>0.043542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.838849</td>\n",
       "      <td>1.079181</td>\n",
       "      <td>1.361728</td>\n",
       "      <td>0.041393</td>\n",
       "      <td>0.301030</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.025306</td>\n",
       "      <td>0.000600</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>-0.122184</td>\n",
       "      <td>...</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.787726</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.165541</td>\n",
       "      <td>2.750655</td>\n",
       "      <td>3.065860</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.241048</td>\n",
       "      <td>0.115237</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>4.00000</td>\n",
       "      <td>12.500000</td>\n",
       "      <td>0.127566</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.072093</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>4.194944</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.881128</td>\n",
       "      <td>3.097306</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.369401</td>\n",
       "      <td>0.142881</td>\n",
       "      <td>0.085146</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>5.00000</td>\n",
       "      <td>13.856000</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.084991</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.091724</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>4.702534</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.418871</td>\n",
       "      <td>3.290758</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.121954</td>\n",
       "      <td>0.146348</td>\n",
       "      <td>0.098837</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>6.00000</td>\n",
       "      <td>19.800000</td>\n",
       "      <td>0.152445</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.114668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6.798393</td>\n",
       "      <td>6.622052</td>\n",
       "      <td>5.887169</td>\n",
       "      <td>4.684449</td>\n",
       "      <td>4.805976</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>3.382502</td>\n",
       "      <td>0.273963</td>\n",
       "      <td>0.195975</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>...</td>\n",
       "      <td>15.00000</td>\n",
       "      <td>500.000000</td>\n",
       "      <td>0.280700</td>\n",
       "      <td>0.229298</td>\n",
       "      <td>0.114102</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>36.490689</td>\n",
       "      <td>16.791555</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.906721</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "count  15628.000000  15628.000000     15628.000000       15628.000000   \n",
       "mean       4.325261      4.487287         3.763063           3.074854   \n",
       "std        0.803837      0.411195         0.777973           0.518909   \n",
       "min        1.838849      1.079181         1.361728           0.041393   \n",
       "25%        3.787726      4.504132         3.165541           2.750655   \n",
       "50%        4.194944      4.504132         3.881128           3.097306   \n",
       "75%        4.702534      4.504132         4.288474           3.418871   \n",
       "max        6.798393      6.622052         5.887169           4.684449   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "count        15628.000000              15628.000000             15628.000000   \n",
       "mean             3.196937                  1.912667                 1.636982   \n",
       "std              0.440543                  4.904279                 0.467608   \n",
       "min              0.301030                  0.990265                 0.025306   \n",
       "25%              3.065860                  0.990265                 1.241048   \n",
       "50%              3.280062                  0.990265                 1.369401   \n",
       "75%              3.290758                  0.990265                 2.121954   \n",
       "max              4.805976                 50.000000                 3.382502   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "count              15628.000000              15628.000000   \n",
       "mean                   0.136264                  0.076520   \n",
       "std                    0.025378                  0.022818   \n",
       "min                    0.000600                  0.052952   \n",
       "25%                    0.115237                  0.052952   \n",
       "50%                    0.142881                  0.085146   \n",
       "75%                    0.146348                  0.098837   \n",
       "max                    0.273963                  0.195975   \n",
       "\n",
       "       forecast_price_pow_p1  ...  num_years_antig       pow_max  \\\n",
       "count           15628.000000  ...      15628.00000  15628.000000   \n",
       "mean               43.767678  ...          5.02310     19.708523   \n",
       "std                 3.787615  ...          1.66229     18.361513   \n",
       "min                -0.122184  ...          1.00000      1.000000   \n",
       "25%                40.606701  ...          4.00000     12.500000   \n",
       "50%                44.311378  ...          5.00000     13.856000   \n",
       "75%                44.311378  ...          6.00000     19.800000   \n",
       "max                59.444710  ...         15.00000    500.000000   \n",
       "\n",
       "       price_p1_var  price_p2_var  price_p3_var  price_p1_fix  price_p2_fix  \\\n",
       "count  15628.000000  15628.000000  15628.000000  15628.000000  15628.000000   \n",
       "mean       0.142776      0.054331      0.029988     43.198620     10.538924   \n",
       "std        0.022686      0.050397      0.035791      5.236536     12.768468   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.127566      0.000000      0.000000     40.565969      0.000000   \n",
       "50%        0.148825      0.084991      0.000000     44.266931      0.000000   \n",
       "75%        0.152445      0.103395      0.071536     44.266931     24.339581   \n",
       "max        0.280700      0.229298      0.114102     59.444710     36.490689   \n",
       "\n",
       "       price_p3_fix         churn  probability_to_churn  \n",
       "count  15628.000000  15628.000000          15628.000000  \n",
       "mean       6.396068      0.097453              0.096452  \n",
       "std        7.769176      0.296584              0.043542  \n",
       "min        0.000000      0.000000              0.000066  \n",
       "25%        0.000000      0.000000              0.072093  \n",
       "50%        0.000000      0.000000              0.091724  \n",
       "75%       16.226389      0.000000              0.114668  \n",
       "max       16.791555      1.000000              0.906721  \n",
       "\n",
       "[8 rows x 26 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" D. Churn Prediction Model Results \"\"\"\n",
    "# Now, we can show each individuals probability to churn based on our model\n",
    "df['probability_to_churn'] = model.predict_proba(df[X_train.columns])[:,1]\n",
    "\n",
    "# merge back id_number column to match id's with probability of churn\n",
    "result = pd.concat([ df, id_number], axis=1)\n",
    "\n",
    "#result.to_excel('churn_probability_by_id.xlsx')\n",
    "result.head()\n",
    "result.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "      <th>churn_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.286767</td>\n",
       "      <td>4.739944</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.374651</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.444045</td>\n",
       "      <td>0.114481</td>\n",
       "      <td>0.098142</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>43.648</td>\n",
       "      <td>0.125976</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>1</td>\n",
       "      <td>0.123745</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.668479</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2.280920</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.237292</td>\n",
       "      <td>0.145711</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>13.800</td>\n",
       "      <td>0.151367</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.075216</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.736397</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>1.689841</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.599009</td>\n",
       "      <td>0.165794</td>\n",
       "      <td>0.087899</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>13.856</td>\n",
       "      <td>0.172468</td>\n",
       "      <td>0.087881</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.098041</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.200029</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2.382089</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.318689</td>\n",
       "      <td>0.146694</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>13.200</td>\n",
       "      <td>0.153133</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.095380</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5.083990</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>4.036071</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.234871</td>\n",
       "      <td>0.110083</td>\n",
       "      <td>0.093746</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>75.000</td>\n",
       "      <td>0.123086</td>\n",
       "      <td>0.100505</td>\n",
       "      <td>0.068646</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0</td>\n",
       "      <td>0.082645</td>\n",
       "      <td>0</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>16038</th>\n",
       "      <td>4.508812</td>\n",
       "      <td>4.680707</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.667360</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.291591</td>\n",
       "      <td>0.138305</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>15.000</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.093501</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16039</th>\n",
       "      <td>3.858778</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>2.260071</td>\n",
       "      <td>2.801191</td>\n",
       "      <td>2.260071</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.161458</td>\n",
       "      <td>0.100167</td>\n",
       "      <td>0.091892</td>\n",
       "      <td>58.995952</td>\n",
       "      <td>...</td>\n",
       "      <td>6.000</td>\n",
       "      <td>0.109747</td>\n",
       "      <td>0.096908</td>\n",
       "      <td>0.070730</td>\n",
       "      <td>58.936774</td>\n",
       "      <td>36.344726</td>\n",
       "      <td>8.334260</td>\n",
       "      <td>1</td>\n",
       "      <td>0.160860</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16040</th>\n",
       "      <td>3.265996</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>2.255273</td>\n",
       "      <td>2.281919</td>\n",
       "      <td>2.255273</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.115943</td>\n",
       "      <td>0.116900</td>\n",
       "      <td>0.100015</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>15.935</td>\n",
       "      <td>0.126098</td>\n",
       "      <td>0.103975</td>\n",
       "      <td>0.070232</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>1</td>\n",
       "      <td>0.158503</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16041</th>\n",
       "      <td>2.120574</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>1.308351</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.912753</td>\n",
       "      <td>0.145711</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>11.000</td>\n",
       "      <td>0.151367</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.151524</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16042</th>\n",
       "      <td>3.941064</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2.882758</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.315970</td>\n",
       "      <td>0.167086</td>\n",
       "      <td>0.088454</td>\n",
       "      <td>45.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>10.392</td>\n",
       "      <td>0.170590</td>\n",
       "      <td>0.087658</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.069080</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>15628 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "1      5.286767      4.739944         4.288474           3.374651   \n",
       "2      3.668479      4.504132         4.288474           2.280920   \n",
       "3      2.736397      4.504132         4.288474           1.689841   \n",
       "4      3.200029      4.504132         4.288474           2.382089   \n",
       "5      5.083990      4.504132         4.093457           4.036071   \n",
       "...         ...           ...              ...                ...   \n",
       "16038  4.508812      4.680707         4.288474           3.667360   \n",
       "16039  3.858778      4.504132         2.260071           2.801191   \n",
       "16040  3.265996      4.504132         2.255273           2.281919   \n",
       "16041  2.120574      4.504132         4.288474           1.308351   \n",
       "16042  3.941064      4.504132         4.288474           2.882758   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "1                3.280062                  0.990265                 0.444045   \n",
       "2                3.280062                  0.990265                 1.237292   \n",
       "3                3.280062                  0.990265                 1.599009   \n",
       "4                3.280062                  0.990265                 1.318689   \n",
       "5                4.093457                  0.990265                 2.234871   \n",
       "...                   ...                       ...                      ...   \n",
       "16038            3.280062                  0.990265                 1.291591   \n",
       "16039            2.260071                  0.990265                 2.161458   \n",
       "16040            2.255273                  0.990265                 2.115943   \n",
       "16041            3.280062                  0.990265                 0.912753   \n",
       "16042            3.280062                  0.990265                 0.315970   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "1                      0.114481                  0.098142   \n",
       "2                      0.145711                  0.052952   \n",
       "3                      0.165794                  0.087899   \n",
       "4                      0.146694                  0.052952   \n",
       "5                      0.110083                  0.093746   \n",
       "...                         ...                       ...   \n",
       "16038                  0.138305                  0.052952   \n",
       "16039                  0.100167                  0.091892   \n",
       "16040                  0.116900                  0.100015   \n",
       "16041                  0.145711                  0.052952   \n",
       "16042                  0.167086                  0.088454   \n",
       "\n",
       "       forecast_price_pow_p1  ...  pow_max  price_p1_var  price_p2_var  \\\n",
       "1                  40.606701  ...   43.648      0.125976      0.103395   \n",
       "2                  44.311378  ...   13.800      0.151367      0.000000   \n",
       "3                  44.311378  ...   13.856      0.172468      0.087881   \n",
       "4                  44.311378  ...   13.200      0.153133      0.000000   \n",
       "5                  40.606701  ...   75.000      0.123086      0.100505   \n",
       "...                      ...  ...      ...           ...           ...   \n",
       "16038              44.311378  ...   15.000      0.148825      0.000000   \n",
       "16039              58.995952  ...    6.000      0.109747      0.096908   \n",
       "16040              40.606701  ...   15.935      0.126098      0.103975   \n",
       "16041              44.311378  ...   11.000      0.151367      0.000000   \n",
       "16042              45.311378  ...   10.392      0.170590      0.087658   \n",
       "\n",
       "       price_p3_var  price_p1_fix  price_p2_fix  price_p3_fix  churn  \\\n",
       "1          0.071536     40.565969     24.339581     16.226389      1   \n",
       "2          0.000000     44.266931      0.000000      0.000000      0   \n",
       "3          0.000000     44.266931      0.000000      0.000000      0   \n",
       "4          0.000000     44.266931      0.000000      0.000000      0   \n",
       "5          0.068646     40.565969     24.339581     16.226389      0   \n",
       "...             ...           ...           ...           ...    ...   \n",
       "16038      0.000000     44.266931      0.000000      0.000000      0   \n",
       "16039      0.070730     58.936774     36.344726      8.334260      1   \n",
       "16040      0.070232     40.565969     24.339581     16.226389      1   \n",
       "16041      0.000000     44.266931      0.000000      0.000000      0   \n",
       "16042      0.000000     44.266931      0.000000      0.000000      0   \n",
       "\n",
       "       probability_to_churn  churn_1  \n",
       "1                  0.123745        0  \n",
       "2                  0.075216        0  \n",
       "3                  0.098041        0  \n",
       "4                  0.095380        0  \n",
       "5                  0.082645        0  \n",
       "...                     ...      ...  \n",
       "16038              0.093501        0  \n",
       "16039              0.160860        1  \n",
       "16040              0.158503        1  \n",
       "16041              0.151524        1  \n",
       "16042              0.069080        0  \n",
       "\n",
       "[15628 rows x 27 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_hat_train = model.predict(X_train)\n",
    "df['churn_1'] = np.where(df['probability_to_churn']>=0.14, 1, 0)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[12863,  1242],\n",
       "       [ 1222,   301]])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" B. Confusion Matrix \"\"\"\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "from sklearn import svm, datasets\n",
    "from sklearn.metrics import confusion_matrix\n",
    "# Pass actual test and predicted target test outcomes to function\n",
    "cnf_matrix = confusion_matrix(df['churn'], df['churn_1'])\n",
    "cnf_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "False positive = 1242, False Negative = 1222, True Positive = 301, True Negative = 12863"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.42857142857142855\n",
      "0.19507453013609852\n",
      "0.004975124378109453\n",
      "0.19763624425476034\n",
      "0.9033754599264118\n",
      "0.8423342718198106\n",
      "0.009836065573770493\n",
      "0.1963470319634703\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score, accuracy_score, f1_score\n",
    "precision_train = precision_score(y_train, y_hat_train)\n",
    "precision_test = precision_score(df['churn'], df['churn_1'])\n",
    "recall_train = recall_score(y_train, y_hat_train)\n",
    "recall_test = recall_score(df['churn'], df['churn_1'])\n",
    "accuracy_train = accuracy_score(y_train, y_hat_train)\n",
    "accuracy_test = accuracy_score(df['churn'], df['churn_1'])\n",
    "f1_train = f1_score(y_train, y_hat_train)\n",
    "f1_test = f1_score(df['churn'], df['churn_1'])\n",
    "print(precision_train)\n",
    "print(precision_test)\n",
    "print(recall_train)\n",
    "print(recall_test)\n",
    "print(accuracy_train)\n",
    "print(accuracy_test)\n",
    "print(f1_train)\n",
    "print(f1_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Although our model accuracy is great, all other measures show lower than expected precision and recall scores."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0    11296\n",
      "1     1206\n",
      "dtype: int64\n",
      "0    0.903535\n",
      "1    0.096465\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "\"\"\"Number of correct predictions in train data \"\"\"\n",
    "\n",
    "# Find residual differences between train data and predicted train data\n",
    "residuals = np.abs(y_train, y_hat_train)\n",
    "# Print the number of times our model was correct ('0') and incorrect ('1')\n",
    "print(pd.Series(residuals).value_counts())\n",
    "# Print normalized amount of times our model was correct (percentage)\n",
    "print(pd.Series(residuals).value_counts(normalize = True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0    2809\n",
      "1     317\n",
      "dtype: int64\n",
      "0    0.898592\n",
      "1    0.101408\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "\"\"\" Number of correct predictions in test data \"\"\"\n",
    "# Find residual differences between test data and predicted test data\n",
    "residual = np.abs(y_test, prediction_test)\n",
    "# Print the number of times our model was correct ('0') and incorrect ('1')\n",
    "print(pd.Series(residual).value_counts())\n",
    "# Print normalized amount of times our model was correct (percentage)\n",
    "print(pd.Series(residual).value_counts(normalize = True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For our case, we will be using the accuracy metric as confirmation. The fact that our prediction accuracy is about the same with both test and train data is a good sign that our model is working well! \n",
    "\n",
    "Both are showing a churn prediction accuracy of around 90%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "      <th>churn_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4.325261</td>\n",
       "      <td>4.487287</td>\n",
       "      <td>3.763063</td>\n",
       "      <td>3.074854</td>\n",
       "      <td>3.196937</td>\n",
       "      <td>1.912667</td>\n",
       "      <td>1.636982</td>\n",
       "      <td>0.136264</td>\n",
       "      <td>0.076520</td>\n",
       "      <td>43.767678</td>\n",
       "      <td>...</td>\n",
       "      <td>19.708523</td>\n",
       "      <td>0.142776</td>\n",
       "      <td>0.054331</td>\n",
       "      <td>0.029988</td>\n",
       "      <td>43.198620</td>\n",
       "      <td>10.538924</td>\n",
       "      <td>6.396068</td>\n",
       "      <td>0.097453</td>\n",
       "      <td>0.096452</td>\n",
       "      <td>0.098733</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.803837</td>\n",
       "      <td>0.411195</td>\n",
       "      <td>0.777973</td>\n",
       "      <td>0.518909</td>\n",
       "      <td>0.440543</td>\n",
       "      <td>4.904279</td>\n",
       "      <td>0.467608</td>\n",
       "      <td>0.025378</td>\n",
       "      <td>0.022818</td>\n",
       "      <td>3.787615</td>\n",
       "      <td>...</td>\n",
       "      <td>18.361513</td>\n",
       "      <td>0.022686</td>\n",
       "      <td>0.050397</td>\n",
       "      <td>0.035791</td>\n",
       "      <td>5.236536</td>\n",
       "      <td>12.768468</td>\n",
       "      <td>7.769176</td>\n",
       "      <td>0.296584</td>\n",
       "      <td>0.043542</td>\n",
       "      <td>0.298313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.838849</td>\n",
       "      <td>1.079181</td>\n",
       "      <td>1.361728</td>\n",
       "      <td>0.041393</td>\n",
       "      <td>0.301030</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.025306</td>\n",
       "      <td>0.000600</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>-0.122184</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000066</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.787726</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.165541</td>\n",
       "      <td>2.750655</td>\n",
       "      <td>3.065860</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.241048</td>\n",
       "      <td>0.115237</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>12.500000</td>\n",
       "      <td>0.127566</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.072093</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>4.194944</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.881128</td>\n",
       "      <td>3.097306</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.369401</td>\n",
       "      <td>0.142881</td>\n",
       "      <td>0.085146</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>13.856000</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.084991</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.091724</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>4.702534</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.418871</td>\n",
       "      <td>3.290758</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.121954</td>\n",
       "      <td>0.146348</td>\n",
       "      <td>0.098837</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>19.800000</td>\n",
       "      <td>0.152445</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.114668</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6.798393</td>\n",
       "      <td>6.622052</td>\n",
       "      <td>5.887169</td>\n",
       "      <td>4.684449</td>\n",
       "      <td>4.805976</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>3.382502</td>\n",
       "      <td>0.273963</td>\n",
       "      <td>0.195975</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>...</td>\n",
       "      <td>500.000000</td>\n",
       "      <td>0.280700</td>\n",
       "      <td>0.229298</td>\n",
       "      <td>0.114102</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>36.490689</td>\n",
       "      <td>16.791555</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.906721</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "count  15628.000000  15628.000000     15628.000000       15628.000000   \n",
       "mean       4.325261      4.487287         3.763063           3.074854   \n",
       "std        0.803837      0.411195         0.777973           0.518909   \n",
       "min        1.838849      1.079181         1.361728           0.041393   \n",
       "25%        3.787726      4.504132         3.165541           2.750655   \n",
       "50%        4.194944      4.504132         3.881128           3.097306   \n",
       "75%        4.702534      4.504132         4.288474           3.418871   \n",
       "max        6.798393      6.622052         5.887169           4.684449   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "count        15628.000000              15628.000000             15628.000000   \n",
       "mean             3.196937                  1.912667                 1.636982   \n",
       "std              0.440543                  4.904279                 0.467608   \n",
       "min              0.301030                  0.990265                 0.025306   \n",
       "25%              3.065860                  0.990265                 1.241048   \n",
       "50%              3.280062                  0.990265                 1.369401   \n",
       "75%              3.290758                  0.990265                 2.121954   \n",
       "max              4.805976                 50.000000                 3.382502   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "count              15628.000000              15628.000000   \n",
       "mean                   0.136264                  0.076520   \n",
       "std                    0.025378                  0.022818   \n",
       "min                    0.000600                  0.052952   \n",
       "25%                    0.115237                  0.052952   \n",
       "50%                    0.142881                  0.085146   \n",
       "75%                    0.146348                  0.098837   \n",
       "max                    0.273963                  0.195975   \n",
       "\n",
       "       forecast_price_pow_p1  ...       pow_max  price_p1_var  price_p2_var  \\\n",
       "count           15628.000000  ...  15628.000000  15628.000000  15628.000000   \n",
       "mean               43.767678  ...     19.708523      0.142776      0.054331   \n",
       "std                 3.787615  ...     18.361513      0.022686      0.050397   \n",
       "min                -0.122184  ...      1.000000      0.000000      0.000000   \n",
       "25%                40.606701  ...     12.500000      0.127566      0.000000   \n",
       "50%                44.311378  ...     13.856000      0.148825      0.084991   \n",
       "75%                44.311378  ...     19.800000      0.152445      0.103395   \n",
       "max                59.444710  ...    500.000000      0.280700      0.229298   \n",
       "\n",
       "       price_p3_var  price_p1_fix  price_p2_fix  price_p3_fix         churn  \\\n",
       "count  15628.000000  15628.000000  15628.000000  15628.000000  15628.000000   \n",
       "mean       0.029988     43.198620     10.538924      6.396068      0.097453   \n",
       "std        0.035791      5.236536     12.768468      7.769176      0.296584   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000     40.565969      0.000000      0.000000      0.000000   \n",
       "50%        0.000000     44.266931      0.000000      0.000000      0.000000   \n",
       "75%        0.071536     44.266931     24.339581     16.226389      0.000000   \n",
       "max        0.114102     59.444710     36.490689     16.791555      1.000000   \n",
       "\n",
       "       probability_to_churn       churn_1  \n",
       "count          15628.000000  15628.000000  \n",
       "mean               0.096452      0.098733  \n",
       "std                0.043542      0.298313  \n",
       "min                0.000066      0.000000  \n",
       "25%                0.072093      0.000000  \n",
       "50%                0.091724      0.000000  \n",
       "75%                0.114668      0.000000  \n",
       "max                0.906721      1.000000  \n",
       "\n",
       "[8 rows x 27 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" D. Churn Prediction Model Results \"\"\"\n",
    "# Now, we can show each individuals probability to churn based on our model\n",
    "df['probability_to_churn'] = model.predict_proba(df[X_train.columns])[:,1]\n",
    "\n",
    "# merge back id_number column to match id's with probability of churn\n",
    "result = pd.concat([ df, id_number], axis=1)\n",
    "\n",
    "#result.to_excel('churn_probability_by_id.xlsx')\n",
    "result.head()\n",
    "result.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "      <th>churn_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.000000</td>\n",
       "      <td>118.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4.011352</td>\n",
       "      <td>4.455240</td>\n",
       "      <td>3.627011</td>\n",
       "      <td>2.922545</td>\n",
       "      <td>3.166054</td>\n",
       "      <td>7.162232</td>\n",
       "      <td>1.840436</td>\n",
       "      <td>0.176897</td>\n",
       "      <td>0.092690</td>\n",
       "      <td>45.809454</td>\n",
       "      <td>...</td>\n",
       "      <td>43.629610</td>\n",
       "      <td>0.165945</td>\n",
       "      <td>0.073098</td>\n",
       "      <td>0.049584</td>\n",
       "      <td>45.081862</td>\n",
       "      <td>15.764857</td>\n",
       "      <td>9.997802</td>\n",
       "      <td>0.262712</td>\n",
       "      <td>0.366999</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.876851</td>\n",
       "      <td>0.330324</td>\n",
       "      <td>0.809465</td>\n",
       "      <td>0.865088</td>\n",
       "      <td>0.647341</td>\n",
       "      <td>10.739011</td>\n",
       "      <td>0.504375</td>\n",
       "      <td>0.063588</td>\n",
       "      <td>0.033823</td>\n",
       "      <td>8.974003</td>\n",
       "      <td>...</td>\n",
       "      <td>39.191141</td>\n",
       "      <td>0.055746</td>\n",
       "      <td>0.055953</td>\n",
       "      <td>0.039541</td>\n",
       "      <td>7.079260</td>\n",
       "      <td>12.393694</td>\n",
       "      <td>7.877867</td>\n",
       "      <td>0.441984</td>\n",
       "      <td>0.132321</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2.068186</td>\n",
       "      <td>2.442480</td>\n",
       "      <td>1.544068</td>\n",
       "      <td>0.372912</td>\n",
       "      <td>1.041393</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.133539</td>\n",
       "      <td>0.112860</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>-0.122184</td>\n",
       "      <td>...</td>\n",
       "      <td>10.350000</td>\n",
       "      <td>0.107070</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>39.818274</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.250500</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.282583</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.227254</td>\n",
       "      <td>2.402060</td>\n",
       "      <td>3.227254</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.289638</td>\n",
       "      <td>0.120372</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>13.287500</td>\n",
       "      <td>0.127408</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.272253</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>4.178214</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.861082</td>\n",
       "      <td>3.016096</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.073369</td>\n",
       "      <td>0.154410</td>\n",
       "      <td>0.100015</td>\n",
       "      <td>41.105201</td>\n",
       "      <td>...</td>\n",
       "      <td>30.375000</td>\n",
       "      <td>0.153159</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.070232</td>\n",
       "      <td>41.063970</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.309924</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>4.651170</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.578037</td>\n",
       "      <td>3.388465</td>\n",
       "      <td>16.747566</td>\n",
       "      <td>2.162333</td>\n",
       "      <td>0.266066</td>\n",
       "      <td>0.125383</td>\n",
       "      <td>57.770878</td>\n",
       "      <td>...</td>\n",
       "      <td>61.500000</td>\n",
       "      <td>0.156255</td>\n",
       "      <td>0.108020</td>\n",
       "      <td>0.074278</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>24.437330</td>\n",
       "      <td>16.291555</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.428105</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6.223976</td>\n",
       "      <td>5.845639</td>\n",
       "      <td>5.308462</td>\n",
       "      <td>4.266765</td>\n",
       "      <td>4.479532</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>2.611691</td>\n",
       "      <td>0.273963</td>\n",
       "      <td>0.168092</td>\n",
       "      <td>59.266378</td>\n",
       "      <td>...</td>\n",
       "      <td>200.000000</td>\n",
       "      <td>0.280700</td>\n",
       "      <td>0.168867</td>\n",
       "      <td>0.101027</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>36.344726</td>\n",
       "      <td>16.724389</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.906721</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "count  118.000000    118.000000       118.000000         118.000000   \n",
       "mean     4.011352      4.455240         3.627011           2.922545   \n",
       "std      0.876851      0.330324         0.809465           0.865088   \n",
       "min      2.068186      2.442480         1.544068           0.372912   \n",
       "25%      3.282583      4.504132         3.227254           2.402060   \n",
       "50%      4.178214      4.504132         3.861082           3.016096   \n",
       "75%      4.651170      4.504132         4.288474           3.578037   \n",
       "max      6.223976      5.845639         5.308462           4.266765   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "count          118.000000                118.000000               118.000000   \n",
       "mean             3.166054                  7.162232                 1.840436   \n",
       "std              0.647341                 10.739011                 0.504375   \n",
       "min              1.041393                  0.990265                 0.133539   \n",
       "25%              3.227254                  0.990265                 1.289638   \n",
       "50%              3.280062                  0.990265                 2.073369   \n",
       "75%              3.388465                 16.747566                 2.162333   \n",
       "max              4.479532                 30.000000                 2.611691   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "count                118.000000                118.000000   \n",
       "mean                   0.176897                  0.092690   \n",
       "std                    0.063588                  0.033823   \n",
       "min                    0.112860                  0.052952   \n",
       "25%                    0.120372                  0.052952   \n",
       "50%                    0.154410                  0.100015   \n",
       "75%                    0.266066                  0.125383   \n",
       "max                    0.273963                  0.168092   \n",
       "\n",
       "       forecast_price_pow_p1  ...     pow_max  price_p1_var  price_p2_var  \\\n",
       "count             118.000000  ...  118.000000    118.000000    118.000000   \n",
       "mean               45.809454  ...   43.629610      0.165945      0.073098   \n",
       "std                 8.974003  ...   39.191141      0.055746      0.055953   \n",
       "min                -0.122184  ...   10.350000      0.107070      0.000000   \n",
       "25%                40.606701  ...   13.287500      0.127408      0.000000   \n",
       "50%                41.105201  ...   30.375000      0.153159      0.103395   \n",
       "75%                57.770878  ...   61.500000      0.156255      0.108020   \n",
       "max                59.266378  ...  200.000000      0.280700      0.168867   \n",
       "\n",
       "       price_p3_var  price_p1_fix  price_p2_fix  price_p3_fix       churn  \\\n",
       "count    118.000000    118.000000    118.000000    118.000000  118.000000   \n",
       "mean       0.049584     45.081862     15.764857      9.997802    0.262712   \n",
       "std        0.039541      7.079260     12.393694      7.877867    0.441984   \n",
       "min        0.000000     39.818274      0.000000      0.000000    0.000000   \n",
       "25%        0.000000     40.565969      0.000000      0.000000    0.000000   \n",
       "50%        0.070232     41.063970     24.339581     16.226389    0.000000   \n",
       "75%        0.074278     44.266931     24.437330     16.291555    1.000000   \n",
       "max        0.101027     59.444710     36.344726     16.724389    1.000000   \n",
       "\n",
       "       probability_to_churn  churn_1  \n",
       "count            118.000000    118.0  \n",
       "mean               0.366999      1.0  \n",
       "std                0.132321      0.0  \n",
       "min                0.250500      1.0  \n",
       "25%                0.272253      1.0  \n",
       "50%                0.309924      1.0  \n",
       "75%                0.428105      1.0  \n",
       "max                0.906721      1.0  \n",
       "\n",
       "[8 rows x 27 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Seperate high likely to churn customers from low likely to churn customers\n",
    "\n",
    "result_high_prob = result[result['probability_to_churn'] >= 0.25]\n",
    "# average probability to churn in this set is 26%\n",
    "\n",
    "result_low_prob = result[result['probability_to_churn'] <= 0.25]\n",
    "# average probability to churn in this set is 10%\n",
    "\n",
    "result_high_prob.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "      <th>churn_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "      <td>15510.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4.327650</td>\n",
       "      <td>4.487530</td>\n",
       "      <td>3.764098</td>\n",
       "      <td>3.076012</td>\n",
       "      <td>3.197172</td>\n",
       "      <td>1.872728</td>\n",
       "      <td>1.635434</td>\n",
       "      <td>0.135955</td>\n",
       "      <td>0.076397</td>\n",
       "      <td>43.752144</td>\n",
       "      <td>...</td>\n",
       "      <td>19.526531</td>\n",
       "      <td>0.142600</td>\n",
       "      <td>0.054188</td>\n",
       "      <td>0.029839</td>\n",
       "      <td>43.184293</td>\n",
       "      <td>10.499165</td>\n",
       "      <td>6.368666</td>\n",
       "      <td>0.096196</td>\n",
       "      <td>0.094394</td>\n",
       "      <td>0.091876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.802817</td>\n",
       "      <td>0.411749</td>\n",
       "      <td>0.777664</td>\n",
       "      <td>0.515259</td>\n",
       "      <td>0.438618</td>\n",
       "      <td>4.811825</td>\n",
       "      <td>0.466995</td>\n",
       "      <td>0.024612</td>\n",
       "      <td>0.022672</td>\n",
       "      <td>3.716945</td>\n",
       "      <td>...</td>\n",
       "      <td>17.992673</td>\n",
       "      <td>0.022159</td>\n",
       "      <td>0.050328</td>\n",
       "      <td>0.035721</td>\n",
       "      <td>5.217727</td>\n",
       "      <td>12.763466</td>\n",
       "      <td>7.762197</td>\n",
       "      <td>0.294869</td>\n",
       "      <td>0.034887</td>\n",
       "      <td>0.288860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.838849</td>\n",
       "      <td>1.079181</td>\n",
       "      <td>1.361728</td>\n",
       "      <td>0.041393</td>\n",
       "      <td>0.301030</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.025306</td>\n",
       "      <td>0.000600</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>32.583108</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000066</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.788928</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.165244</td>\n",
       "      <td>2.753137</td>\n",
       "      <td>3.065206</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.240799</td>\n",
       "      <td>0.115237</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>12.500000</td>\n",
       "      <td>0.127566</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.071932</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>4.194944</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.881556</td>\n",
       "      <td>3.097519</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.360688</td>\n",
       "      <td>0.142881</td>\n",
       "      <td>0.085067</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>13.856000</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.084769</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.091447</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>4.702534</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.417563</td>\n",
       "      <td>3.290257</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.121954</td>\n",
       "      <td>0.146343</td>\n",
       "      <td>0.098837</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>19.800000</td>\n",
       "      <td>0.152445</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.114046</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6.798393</td>\n",
       "      <td>6.622052</td>\n",
       "      <td>5.887169</td>\n",
       "      <td>4.684449</td>\n",
       "      <td>4.805976</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>3.382502</td>\n",
       "      <td>0.273963</td>\n",
       "      <td>0.195975</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>...</td>\n",
       "      <td>500.000000</td>\n",
       "      <td>0.280700</td>\n",
       "      <td>0.229298</td>\n",
       "      <td>0.114102</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>36.490689</td>\n",
       "      <td>16.791555</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.249248</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "count  15510.000000  15510.000000     15510.000000       15510.000000   \n",
       "mean       4.327650      4.487530         3.764098           3.076012   \n",
       "std        0.802817      0.411749         0.777664           0.515259   \n",
       "min        1.838849      1.079181         1.361728           0.041393   \n",
       "25%        3.788928      4.504132         3.165244           2.753137   \n",
       "50%        4.194944      4.504132         3.881556           3.097519   \n",
       "75%        4.702534      4.504132         4.288474           3.417563   \n",
       "max        6.798393      6.622052         5.887169           4.684449   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "count        15510.000000              15510.000000             15510.000000   \n",
       "mean             3.197172                  1.872728                 1.635434   \n",
       "std              0.438618                  4.811825                 0.466995   \n",
       "min              0.301030                  0.990265                 0.025306   \n",
       "25%              3.065206                  0.990265                 1.240799   \n",
       "50%              3.280062                  0.990265                 1.360688   \n",
       "75%              3.290257                  0.990265                 2.121954   \n",
       "max              4.805976                 50.000000                 3.382502   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "count              15510.000000              15510.000000   \n",
       "mean                   0.135955                  0.076397   \n",
       "std                    0.024612                  0.022672   \n",
       "min                    0.000600                  0.052952   \n",
       "25%                    0.115237                  0.052952   \n",
       "50%                    0.142881                  0.085067   \n",
       "75%                    0.146343                  0.098837   \n",
       "max                    0.273963                  0.195975   \n",
       "\n",
       "       forecast_price_pow_p1  ...       pow_max  price_p1_var  price_p2_var  \\\n",
       "count           15510.000000  ...  15510.000000  15510.000000  15510.000000   \n",
       "mean               43.752144  ...     19.526531      0.142600      0.054188   \n",
       "std                 3.716945  ...     17.992673      0.022159      0.050328   \n",
       "min                32.583108  ...      1.000000      0.000000      0.000000   \n",
       "25%                40.606701  ...     12.500000      0.127566      0.000000   \n",
       "50%                44.311378  ...     13.856000      0.148825      0.084769   \n",
       "75%                44.311378  ...     19.800000      0.152445      0.103395   \n",
       "max                59.444710  ...    500.000000      0.280700      0.229298   \n",
       "\n",
       "       price_p3_var  price_p1_fix  price_p2_fix  price_p3_fix         churn  \\\n",
       "count  15510.000000  15510.000000  15510.000000  15510.000000  15510.000000   \n",
       "mean       0.029839     43.184293     10.499165      6.368666      0.096196   \n",
       "std        0.035721      5.217727     12.763466      7.762197      0.294869   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000     40.565969      0.000000      0.000000      0.000000   \n",
       "50%        0.000000     44.266931      0.000000      0.000000      0.000000   \n",
       "75%        0.071536     44.266931     24.339581     16.226389      0.000000   \n",
       "max        0.114102     59.444710     36.490689     16.791555      1.000000   \n",
       "\n",
       "       probability_to_churn       churn_1  \n",
       "count          15510.000000  15510.000000  \n",
       "mean               0.094394      0.091876  \n",
       "std                0.034887      0.288860  \n",
       "min                0.000066      0.000000  \n",
       "25%                0.071932      0.000000  \n",
       "50%                0.091447      0.000000  \n",
       "75%                0.114046      0.000000  \n",
       "max                0.249248      1.000000  \n",
       "\n",
       "[8 rows x 27 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_low_prob.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 10. Discount Hypothesis Test\n",
    "    A. Identify Variable weights\n",
    "    B. Provide 20% discount\n",
    "    C. Run Model with 20% discount\n",
    "    D. Hypothesis Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "forecast_cons_year          0.749778\n",
       "price_p1_fix                0.057501\n",
       "price_p2_fix                0.023279\n",
       "margin_gross_pow_ele        0.009916\n",
       "margin_net_pow_ele          0.002312\n",
       "net_margin                  0.001095\n",
       "pow_max                    -0.006307\n",
       "forecast_discount_energy   -0.009654\n",
       "price_p3_fix               -0.026570\n",
       "nb_prod_act                -0.037065\n",
       "cons_last_month            -0.040804\n",
       "price_p3_var               -0.041858\n",
       "forecast_meter_rent_12m    -0.043158\n",
       "forecast_cons_12m          -0.047956\n",
       "forecast_price_pow_p1      -0.068061\n",
       "price_p1_var               -0.144565\n",
       "num_years_antig            -0.158899\n",
       "price_p2_var               -0.175382\n",
       "cons_12m                   -0.188309\n",
       "cons_gas_12m               -0.238936\n",
       "forecast_price_energy_p1   -0.264569\n",
       "forecast_price_energy_p2   -0.288296\n",
       "has_gas                    -0.306796\n",
       "imp_cons                   -0.732759\n",
       "dtype: float64"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" A. Identify Variable Weights \"\"\"\n",
    "# To get the weights of all the variables\n",
    "weights = pd.Series(model.coef_[0], index=X.columns.values)\n",
    "weights.sort_values(ascending = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The top five positive variable weights are imp_cons, forecast_meter_rent_12m, price_p2_fix, forecast_cons_year, and margin_gross_pow_ele. These variables promote customer churn and should be looked at further."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will now focus on reducing variable 'price_p2_fix' by 20% according to our client PowerCo's reommendation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>net_margin</th>\n",
       "      <th>num_years_antig</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.286767</td>\n",
       "      <td>4.739944</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.374651</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.444045</td>\n",
       "      <td>0.114481</td>\n",
       "      <td>0.098142</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>678.99</td>\n",
       "      <td>3</td>\n",
       "      <td>43.648</td>\n",
       "      <td>0.125976</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>19.471665</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.668479</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2.280920</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.237292</td>\n",
       "      <td>0.145711</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>18.89</td>\n",
       "      <td>6</td>\n",
       "      <td>13.800</td>\n",
       "      <td>0.151367</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.736397</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>1.689841</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.599009</td>\n",
       "      <td>0.165794</td>\n",
       "      <td>0.087899</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>6.60</td>\n",
       "      <td>6</td>\n",
       "      <td>13.856</td>\n",
       "      <td>0.172468</td>\n",
       "      <td>0.087881</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.200029</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2.382089</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.318689</td>\n",
       "      <td>0.146694</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>25.46</td>\n",
       "      <td>6</td>\n",
       "      <td>13.200</td>\n",
       "      <td>0.153133</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5.083990</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>4.036071</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.234871</td>\n",
       "      <td>0.110083</td>\n",
       "      <td>0.093746</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>823.18</td>\n",
       "      <td>6</td>\n",
       "      <td>75.000</td>\n",
       "      <td>0.123086</td>\n",
       "      <td>0.100505</td>\n",
       "      <td>0.068646</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>19.471665</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "1  5.286767      4.739944         4.288474           3.374651   \n",
       "2  3.668479      4.504132         4.288474           2.280920   \n",
       "3  2.736397      4.504132         4.288474           1.689841   \n",
       "4  3.200029      4.504132         4.288474           2.382089   \n",
       "5  5.083990      4.504132         4.093457           4.036071   \n",
       "\n",
       "   forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "1            3.280062                  0.990265                 0.444045   \n",
       "2            3.280062                  0.990265                 1.237292   \n",
       "3            3.280062                  0.990265                 1.599009   \n",
       "4            3.280062                  0.990265                 1.318689   \n",
       "5            4.093457                  0.990265                 2.234871   \n",
       "\n",
       "   forecast_price_energy_p1  forecast_price_energy_p2  forecast_price_pow_p1  \\\n",
       "1                  0.114481                  0.098142              40.606701   \n",
       "2                  0.145711                  0.052952              44.311378   \n",
       "3                  0.165794                  0.087899              44.311378   \n",
       "4                  0.146694                  0.052952              44.311378   \n",
       "5                  0.110083                  0.093746              40.606701   \n",
       "\n",
       "   ...  net_margin  num_years_antig  pow_max  price_p1_var  price_p2_var  \\\n",
       "1  ...      678.99                3   43.648      0.125976      0.103395   \n",
       "2  ...       18.89                6   13.800      0.151367      0.000000   \n",
       "3  ...        6.60                6   13.856      0.172468      0.087881   \n",
       "4  ...       25.46                6   13.200      0.153133      0.000000   \n",
       "5  ...      823.18                6   75.000      0.123086      0.100505   \n",
       "\n",
       "   price_p3_var  price_p1_fix  price_p2_fix  price_p3_fix  churn  \n",
       "1      0.071536     40.565969     19.471665     16.226389      1  \n",
       "2      0.000000     44.266931      0.000000      0.000000      0  \n",
       "3      0.000000     44.266931      0.000000      0.000000      0  \n",
       "4      0.000000     44.266931      0.000000      0.000000      0  \n",
       "5      0.068646     40.565969     19.471665     16.226389      0  \n",
       "\n",
       "[5 rows x 25 columns]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" B. Provide 20% Discount \"\"\"\n",
    "# apply the 20% discount onto df's 'price_p2_fix' column and redo feature scaling to run the model and\n",
    "# look for any noticible change in customer churn amount.\n",
    "\n",
    "df_new = df.drop(labels = [\"probability_to_churn\", \"churn_1\"], axis = 1)\n",
    "df_new['price_p2_fix'] = df_new['price_p2_fix'].astype(float)/1.25\n",
    "#df_new['price_p3_fix'] = df_new['price_p3_fix'].astype(float)/1.25\n",
    "#comment or uncomment the above line if you would like to test the scenarios listed\n",
    "#in section 10D\n",
    "\n",
    "df_new.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\" C. Run Model with 20% discount \"\"\"\n",
    "# Create 'test' data set by splitting df_new into test/training data\n",
    "# Now, normalize the new set\n",
    "Y_new = df_new[\"churn\"].values\n",
    "X_new = df_new.drop(labels = [\"churn\"],axis = 1)\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "X_train_new, X_test_new, y_train_new, y_test_new = train_test_split(X_new, Y_new, test_size=0.3, random_state=101)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>num_years_antig</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.286767</td>\n",
       "      <td>4.739944</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.374651</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.444045</td>\n",
       "      <td>0.114481</td>\n",
       "      <td>0.098142</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>43.648</td>\n",
       "      <td>0.125976</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>19.471665</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>1</td>\n",
       "      <td>0.111972</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.668479</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2.280920</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.237292</td>\n",
       "      <td>0.145711</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>13.800</td>\n",
       "      <td>0.151367</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.075216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.736397</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>1.689841</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.599009</td>\n",
       "      <td>0.165794</td>\n",
       "      <td>0.087899</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>13.856</td>\n",
       "      <td>0.172468</td>\n",
       "      <td>0.087881</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.098041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.200029</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>2.382089</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.318689</td>\n",
       "      <td>0.146694</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>13.200</td>\n",
       "      <td>0.153133</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.095380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5.083990</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>4.036071</td>\n",
       "      <td>4.093457</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.234871</td>\n",
       "      <td>0.110083</td>\n",
       "      <td>0.093746</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>75.000</td>\n",
       "      <td>0.123086</td>\n",
       "      <td>0.100505</td>\n",
       "      <td>0.068646</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>19.471665</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0</td>\n",
       "      <td>0.074450</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "1  5.286767      4.739944         4.288474           3.374651   \n",
       "2  3.668479      4.504132         4.288474           2.280920   \n",
       "3  2.736397      4.504132         4.288474           1.689841   \n",
       "4  3.200029      4.504132         4.288474           2.382089   \n",
       "5  5.083990      4.504132         4.093457           4.036071   \n",
       "\n",
       "   forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "1            3.280062                  0.990265                 0.444045   \n",
       "2            3.280062                  0.990265                 1.237292   \n",
       "3            3.280062                  0.990265                 1.599009   \n",
       "4            3.280062                  0.990265                 1.318689   \n",
       "5            4.093457                  0.990265                 2.234871   \n",
       "\n",
       "   forecast_price_energy_p1  forecast_price_energy_p2  forecast_price_pow_p1  \\\n",
       "1                  0.114481                  0.098142              40.606701   \n",
       "2                  0.145711                  0.052952              44.311378   \n",
       "3                  0.165794                  0.087899              44.311378   \n",
       "4                  0.146694                  0.052952              44.311378   \n",
       "5                  0.110083                  0.093746              40.606701   \n",
       "\n",
       "   ...  num_years_antig  pow_max  price_p1_var  price_p2_var  price_p3_var  \\\n",
       "1  ...                3   43.648      0.125976      0.103395      0.071536   \n",
       "2  ...                6   13.800      0.151367      0.000000      0.000000   \n",
       "3  ...                6   13.856      0.172468      0.087881      0.000000   \n",
       "4  ...                6   13.200      0.153133      0.000000      0.000000   \n",
       "5  ...                6   75.000      0.123086      0.100505      0.068646   \n",
       "\n",
       "   price_p1_fix  price_p2_fix  price_p3_fix  churn  probability_to_churn  \n",
       "1     40.565969     19.471665     16.226389      1              0.111972  \n",
       "2     44.266931      0.000000      0.000000      0              0.075216  \n",
       "3     44.266931      0.000000      0.000000      0              0.098041  \n",
       "4     44.266931      0.000000      0.000000      0              0.095380  \n",
       "5     40.565969     19.471665     16.226389      0              0.074450  \n",
       "\n",
       "[5 rows x 26 columns]"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_new['probability_to_churn'] = model.predict_proba(df_new[X_train.columns])[:,1]\n",
    "df_new.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "y_hat_train_new = model.predict(X_train_new)\n",
    "y_hat_test_new = model.predict(X_test_new)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[4230,    1],\n",
       "       [ 455,    3]])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\" D. New 20% Reduced Cost Results \"\"\"\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "from sklearn import svm, datasets\n",
    "from sklearn.metrics import confusion_matrix\n",
    "# Pass actual test and predicted target test outcomes to function\n",
    "cnf_matrix_new = confusion_matrix(y_test_new, y_hat_test_new)\n",
    "cnf_matrix_new"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>num_years_antig</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>15628.00000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4.325261</td>\n",
       "      <td>4.487287</td>\n",
       "      <td>3.763063</td>\n",
       "      <td>3.074854</td>\n",
       "      <td>3.196937</td>\n",
       "      <td>1.912667</td>\n",
       "      <td>1.636982</td>\n",
       "      <td>0.136264</td>\n",
       "      <td>0.076520</td>\n",
       "      <td>43.767678</td>\n",
       "      <td>...</td>\n",
       "      <td>5.02310</td>\n",
       "      <td>19.708523</td>\n",
       "      <td>0.142776</td>\n",
       "      <td>0.054331</td>\n",
       "      <td>0.029988</td>\n",
       "      <td>43.198620</td>\n",
       "      <td>8.431139</td>\n",
       "      <td>6.396068</td>\n",
       "      <td>0.097453</td>\n",
       "      <td>0.092073</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.803837</td>\n",
       "      <td>0.411195</td>\n",
       "      <td>0.777973</td>\n",
       "      <td>0.518909</td>\n",
       "      <td>0.440543</td>\n",
       "      <td>4.904279</td>\n",
       "      <td>0.467608</td>\n",
       "      <td>0.025378</td>\n",
       "      <td>0.022818</td>\n",
       "      <td>3.787615</td>\n",
       "      <td>...</td>\n",
       "      <td>1.66229</td>\n",
       "      <td>18.361513</td>\n",
       "      <td>0.022686</td>\n",
       "      <td>0.050397</td>\n",
       "      <td>0.035791</td>\n",
       "      <td>5.236536</td>\n",
       "      <td>10.214775</td>\n",
       "      <td>7.769176</td>\n",
       "      <td>0.296584</td>\n",
       "      <td>0.040958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.838849</td>\n",
       "      <td>1.079181</td>\n",
       "      <td>1.361728</td>\n",
       "      <td>0.041393</td>\n",
       "      <td>0.301030</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.025306</td>\n",
       "      <td>0.000600</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>-0.122184</td>\n",
       "      <td>...</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.787726</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.165541</td>\n",
       "      <td>2.750655</td>\n",
       "      <td>3.065860</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.241048</td>\n",
       "      <td>0.115237</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>4.00000</td>\n",
       "      <td>12.500000</td>\n",
       "      <td>0.127566</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.070051</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>4.194944</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.881128</td>\n",
       "      <td>3.097306</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.369401</td>\n",
       "      <td>0.142881</td>\n",
       "      <td>0.085146</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>5.00000</td>\n",
       "      <td>13.856000</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.084991</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.088431</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>4.702534</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.418871</td>\n",
       "      <td>3.290758</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.121954</td>\n",
       "      <td>0.146348</td>\n",
       "      <td>0.098837</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>6.00000</td>\n",
       "      <td>19.800000</td>\n",
       "      <td>0.152445</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>19.471665</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.109806</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6.798393</td>\n",
       "      <td>6.622052</td>\n",
       "      <td>5.887169</td>\n",
       "      <td>4.684449</td>\n",
       "      <td>4.805976</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>3.382502</td>\n",
       "      <td>0.273963</td>\n",
       "      <td>0.195975</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>...</td>\n",
       "      <td>15.00000</td>\n",
       "      <td>500.000000</td>\n",
       "      <td>0.280700</td>\n",
       "      <td>0.229298</td>\n",
       "      <td>0.114102</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>29.192551</td>\n",
       "      <td>16.791555</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.896685</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "count  15628.000000  15628.000000     15628.000000       15628.000000   \n",
       "mean       4.325261      4.487287         3.763063           3.074854   \n",
       "std        0.803837      0.411195         0.777973           0.518909   \n",
       "min        1.838849      1.079181         1.361728           0.041393   \n",
       "25%        3.787726      4.504132         3.165541           2.750655   \n",
       "50%        4.194944      4.504132         3.881128           3.097306   \n",
       "75%        4.702534      4.504132         4.288474           3.418871   \n",
       "max        6.798393      6.622052         5.887169           4.684449   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "count        15628.000000              15628.000000             15628.000000   \n",
       "mean             3.196937                  1.912667                 1.636982   \n",
       "std              0.440543                  4.904279                 0.467608   \n",
       "min              0.301030                  0.990265                 0.025306   \n",
       "25%              3.065860                  0.990265                 1.241048   \n",
       "50%              3.280062                  0.990265                 1.369401   \n",
       "75%              3.290758                  0.990265                 2.121954   \n",
       "max              4.805976                 50.000000                 3.382502   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "count              15628.000000              15628.000000   \n",
       "mean                   0.136264                  0.076520   \n",
       "std                    0.025378                  0.022818   \n",
       "min                    0.000600                  0.052952   \n",
       "25%                    0.115237                  0.052952   \n",
       "50%                    0.142881                  0.085146   \n",
       "75%                    0.146348                  0.098837   \n",
       "max                    0.273963                  0.195975   \n",
       "\n",
       "       forecast_price_pow_p1  ...  num_years_antig       pow_max  \\\n",
       "count           15628.000000  ...      15628.00000  15628.000000   \n",
       "mean               43.767678  ...          5.02310     19.708523   \n",
       "std                 3.787615  ...          1.66229     18.361513   \n",
       "min                -0.122184  ...          1.00000      1.000000   \n",
       "25%                40.606701  ...          4.00000     12.500000   \n",
       "50%                44.311378  ...          5.00000     13.856000   \n",
       "75%                44.311378  ...          6.00000     19.800000   \n",
       "max                59.444710  ...         15.00000    500.000000   \n",
       "\n",
       "       price_p1_var  price_p2_var  price_p3_var  price_p1_fix  price_p2_fix  \\\n",
       "count  15628.000000  15628.000000  15628.000000  15628.000000  15628.000000   \n",
       "mean       0.142776      0.054331      0.029988     43.198620      8.431139   \n",
       "std        0.022686      0.050397      0.035791      5.236536     10.214775   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.127566      0.000000      0.000000     40.565969      0.000000   \n",
       "50%        0.148825      0.084991      0.000000     44.266931      0.000000   \n",
       "75%        0.152445      0.103395      0.071536     44.266931     19.471665   \n",
       "max        0.280700      0.229298      0.114102     59.444710     29.192551   \n",
       "\n",
       "       price_p3_fix         churn  probability_to_churn  \n",
       "count  15628.000000  15628.000000          15628.000000  \n",
       "mean       6.396068      0.097453              0.092073  \n",
       "std        7.769176      0.296584              0.040958  \n",
       "min        0.000000      0.000000              0.000060  \n",
       "25%        0.000000      0.000000              0.070051  \n",
       "50%        0.000000      0.000000              0.088431  \n",
       "75%       16.226389      0.000000              0.109806  \n",
       "max       16.791555      1.000000              0.896685  \n",
       "\n",
       "[8 rows x 26 columns]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_new.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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>cons_12m</th>\n",
       "      <th>cons_gas_12m</th>\n",
       "      <th>cons_last_month</th>\n",
       "      <th>forecast_cons_12m</th>\n",
       "      <th>forecast_cons_year</th>\n",
       "      <th>forecast_discount_energy</th>\n",
       "      <th>forecast_meter_rent_12m</th>\n",
       "      <th>forecast_price_energy_p1</th>\n",
       "      <th>forecast_price_energy_p2</th>\n",
       "      <th>forecast_price_pow_p1</th>\n",
       "      <th>...</th>\n",
       "      <th>pow_max</th>\n",
       "      <th>price_p1_var</th>\n",
       "      <th>price_p2_var</th>\n",
       "      <th>price_p3_var</th>\n",
       "      <th>price_p1_fix</th>\n",
       "      <th>price_p2_fix</th>\n",
       "      <th>price_p3_fix</th>\n",
       "      <th>churn</th>\n",
       "      <th>probability_to_churn</th>\n",
       "      <th>churn_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "      <td>15628.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>4.325261</td>\n",
       "      <td>4.487287</td>\n",
       "      <td>3.763063</td>\n",
       "      <td>3.074854</td>\n",
       "      <td>3.196937</td>\n",
       "      <td>1.912667</td>\n",
       "      <td>1.636982</td>\n",
       "      <td>0.136264</td>\n",
       "      <td>0.076520</td>\n",
       "      <td>43.767678</td>\n",
       "      <td>...</td>\n",
       "      <td>19.708523</td>\n",
       "      <td>0.142776</td>\n",
       "      <td>0.054331</td>\n",
       "      <td>0.029988</td>\n",
       "      <td>43.198620</td>\n",
       "      <td>10.538924</td>\n",
       "      <td>6.396068</td>\n",
       "      <td>0.097453</td>\n",
       "      <td>0.096452</td>\n",
       "      <td>0.098733</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.803837</td>\n",
       "      <td>0.411195</td>\n",
       "      <td>0.777973</td>\n",
       "      <td>0.518909</td>\n",
       "      <td>0.440543</td>\n",
       "      <td>4.904279</td>\n",
       "      <td>0.467608</td>\n",
       "      <td>0.025378</td>\n",
       "      <td>0.022818</td>\n",
       "      <td>3.787615</td>\n",
       "      <td>...</td>\n",
       "      <td>18.361513</td>\n",
       "      <td>0.022686</td>\n",
       "      <td>0.050397</td>\n",
       "      <td>0.035791</td>\n",
       "      <td>5.236536</td>\n",
       "      <td>12.768468</td>\n",
       "      <td>7.769176</td>\n",
       "      <td>0.296584</td>\n",
       "      <td>0.043542</td>\n",
       "      <td>0.298313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.838849</td>\n",
       "      <td>1.079181</td>\n",
       "      <td>1.361728</td>\n",
       "      <td>0.041393</td>\n",
       "      <td>0.301030</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>0.025306</td>\n",
       "      <td>0.000600</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>-0.122184</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000066</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.787726</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.165541</td>\n",
       "      <td>2.750655</td>\n",
       "      <td>3.065860</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.241048</td>\n",
       "      <td>0.115237</td>\n",
       "      <td>0.052952</td>\n",
       "      <td>40.606701</td>\n",
       "      <td>...</td>\n",
       "      <td>12.500000</td>\n",
       "      <td>0.127566</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.565969</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.072093</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>4.194944</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>3.881128</td>\n",
       "      <td>3.097306</td>\n",
       "      <td>3.280062</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>1.369401</td>\n",
       "      <td>0.142881</td>\n",
       "      <td>0.085146</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>13.856000</td>\n",
       "      <td>0.148825</td>\n",
       "      <td>0.084991</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.091724</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>4.702534</td>\n",
       "      <td>4.504132</td>\n",
       "      <td>4.288474</td>\n",
       "      <td>3.418871</td>\n",
       "      <td>3.290758</td>\n",
       "      <td>0.990265</td>\n",
       "      <td>2.121954</td>\n",
       "      <td>0.146348</td>\n",
       "      <td>0.098837</td>\n",
       "      <td>44.311378</td>\n",
       "      <td>...</td>\n",
       "      <td>19.800000</td>\n",
       "      <td>0.152445</td>\n",
       "      <td>0.103395</td>\n",
       "      <td>0.071536</td>\n",
       "      <td>44.266931</td>\n",
       "      <td>24.339581</td>\n",
       "      <td>16.226389</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.114668</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6.798393</td>\n",
       "      <td>6.622052</td>\n",
       "      <td>5.887169</td>\n",
       "      <td>4.684449</td>\n",
       "      <td>4.805976</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>3.382502</td>\n",
       "      <td>0.273963</td>\n",
       "      <td>0.195975</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>...</td>\n",
       "      <td>500.000000</td>\n",
       "      <td>0.280700</td>\n",
       "      <td>0.229298</td>\n",
       "      <td>0.114102</td>\n",
       "      <td>59.444710</td>\n",
       "      <td>36.490689</td>\n",
       "      <td>16.791555</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.906721</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 27 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           cons_12m  cons_gas_12m  cons_last_month  forecast_cons_12m  \\\n",
       "count  15628.000000  15628.000000     15628.000000       15628.000000   \n",
       "mean       4.325261      4.487287         3.763063           3.074854   \n",
       "std        0.803837      0.411195         0.777973           0.518909   \n",
       "min        1.838849      1.079181         1.361728           0.041393   \n",
       "25%        3.787726      4.504132         3.165541           2.750655   \n",
       "50%        4.194944      4.504132         3.881128           3.097306   \n",
       "75%        4.702534      4.504132         4.288474           3.418871   \n",
       "max        6.798393      6.622052         5.887169           4.684449   \n",
       "\n",
       "       forecast_cons_year  forecast_discount_energy  forecast_meter_rent_12m  \\\n",
       "count        15628.000000              15628.000000             15628.000000   \n",
       "mean             3.196937                  1.912667                 1.636982   \n",
       "std              0.440543                  4.904279                 0.467608   \n",
       "min              0.301030                  0.990265                 0.025306   \n",
       "25%              3.065860                  0.990265                 1.241048   \n",
       "50%              3.280062                  0.990265                 1.369401   \n",
       "75%              3.290758                  0.990265                 2.121954   \n",
       "max              4.805976                 50.000000                 3.382502   \n",
       "\n",
       "       forecast_price_energy_p1  forecast_price_energy_p2  \\\n",
       "count              15628.000000              15628.000000   \n",
       "mean                   0.136264                  0.076520   \n",
       "std                    0.025378                  0.022818   \n",
       "min                    0.000600                  0.052952   \n",
       "25%                    0.115237                  0.052952   \n",
       "50%                    0.142881                  0.085146   \n",
       "75%                    0.146348                  0.098837   \n",
       "max                    0.273963                  0.195975   \n",
       "\n",
       "       forecast_price_pow_p1  ...       pow_max  price_p1_var  price_p2_var  \\\n",
       "count           15628.000000  ...  15628.000000  15628.000000  15628.000000   \n",
       "mean               43.767678  ...     19.708523      0.142776      0.054331   \n",
       "std                 3.787615  ...     18.361513      0.022686      0.050397   \n",
       "min                -0.122184  ...      1.000000      0.000000      0.000000   \n",
       "25%                40.606701  ...     12.500000      0.127566      0.000000   \n",
       "50%                44.311378  ...     13.856000      0.148825      0.084991   \n",
       "75%                44.311378  ...     19.800000      0.152445      0.103395   \n",
       "max                59.444710  ...    500.000000      0.280700      0.229298   \n",
       "\n",
       "       price_p3_var  price_p1_fix  price_p2_fix  price_p3_fix         churn  \\\n",
       "count  15628.000000  15628.000000  15628.000000  15628.000000  15628.000000   \n",
       "mean       0.029988     43.198620     10.538924      6.396068      0.097453   \n",
       "std        0.035791      5.236536     12.768468      7.769176      0.296584   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000     40.565969      0.000000      0.000000      0.000000   \n",
       "50%        0.000000     44.266931      0.000000      0.000000      0.000000   \n",
       "75%        0.071536     44.266931     24.339581     16.226389      0.000000   \n",
       "max        0.114102     59.444710     36.490689     16.791555      1.000000   \n",
       "\n",
       "       probability_to_churn       churn_1  \n",
       "count          15628.000000  15628.000000  \n",
       "mean               0.096452      0.098733  \n",
       "std                0.043542      0.298313  \n",
       "min                0.000066      0.000000  \n",
       "25%                0.072093      0.000000  \n",
       "50%                0.091724      0.000000  \n",
       "75%                0.114668      0.000000  \n",
       "max                0.906721      1.000000  \n",
       "\n",
       "[8 rows x 27 columns]"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Average probability to churn goes down from 9.8% to 9.2% by reducing 'price_p2_fix' \n",
    "# by 20%. \n",
    "\n",
    "# (The following scenarios are not shown in this notebook)\n",
    "# 1. Churn probability actually goes up to 10.5% when reducing 'price_p3_fix' by 20% and\n",
    "# leaving 'price_p2_fix' alone.\n",
    "\n",
    "# 2. Average probability to churn goes down from 10.25% to 9.9% by reducing 'price_p2_fix' \n",
    "# and 'price_p3_fix' by 20%."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}