CustomerChurnAnalysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Customer Churn Analysis - Jupyter Notebook\n",
    "## Outline:\n",
    "*   I. Introduction\n",
    "*  II. Exploratory Data Analysis\n",
    "* III. Data Preprocessing\n",
    "*  IV. Modeling\n",
    "    * i. Random Forest Classifier\n",
    "    * ii. Logistic Regression Classifier\n",
    "    * iii. Support Vector Machine Classifier\n",
    "    * iv. SMOTE implementation on all three previous models\n",
    "*   V. Conclusions "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## I. Introduction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset we are experimenting with was found [here](https://www.kaggle.com/blastchar/telco-customer-churn). To gain insight on what we are working with, I've copied and pasted the bulk of the description from the link:\n",
    "\n",
    "> **Context**\n",
    ">\n",
    "> \"Predict behavior to retain customers. You can analyze all relevant customer data and develop focused customer retention programs.\" \\[IBM Sample Data > Sets\\]\n",
    ">\n",
    "> **Content**\n",
    ">\n",
    "> Each row represents a customer, each column contains customer’s attributes described on the column Metadata.\n",
    ">\n",
    "> The data set includes information about:\n",
    ">\n",
    "> * Customers who left within the last month – the column is called Churn\n",
    "> * Services that each customer has signed up for – phone, multiple lines, internet, online security, online backup, device protection, tech support, and streaming TV and movies\n",
    "> * Customer account information – how long they’ve been a customer, contract, payment method, paperless billing, monthly charges, and total charges\n",
    "> * Demographic info about customers – gender, age range, and if they have partners and dependents\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The link above doesn't tell us much about where the data is gathered from, but based on the description of the variables/columns, we can assume that this data is from a telecommuncations company similar to that of Verizon or Comcast.\n",
    "\n",
    "*Description of columns will be shown in the next section*\n",
    "\n",
    "As stated in the **Content** section, \"Churn\" refers to whether or not the customer left the company within the last month or stayed with the company. This will be our target variable.\n",
    "\n",
    "We're going to investigate this problem using **Random Forest**, **Logistic Regression**, and **Support Vector Machine** to see what information we can find. We can also use these methods to raise new questions and strategies for customer retention."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Libraries we'll use:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* More libraries will be imported later in the project. Here below is what we need to get started."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "col_scheme = [\"#4374B3\", \"#FF0B04\"] # Color scheme for our plots"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## II. Exploratory Data Analysis (and some cleaning)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's get started by getting this dataset loaded into our notebook for a small glimpse of the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customerID</th>\n",
       "      <th>gender</th>\n",
       "      <th>SeniorCitizen</th>\n",
       "      <th>Partner</th>\n",
       "      <th>Dependents</th>\n",
       "      <th>tenure</th>\n",
       "      <th>PhoneService</th>\n",
       "      <th>MultipleLines</th>\n",
       "      <th>InternetService</th>\n",
       "      <th>OnlineSecurity</th>\n",
       "      <th>...</th>\n",
       "      <th>DeviceProtection</th>\n",
       "      <th>TechSupport</th>\n",
       "      <th>StreamingTV</th>\n",
       "      <th>StreamingMovies</th>\n",
       "      <th>Contract</th>\n",
       "      <th>PaperlessBilling</th>\n",
       "      <th>PaymentMethod</th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>Churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7590-VHVEG</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>No</td>\n",
       "      <td>No phone service</td>\n",
       "      <td>DSL</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>29.85</td>\n",
       "      <td>29.85</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>34</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>56.95</td>\n",
       "      <td>1889.5</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>2</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>53.85</td>\n",
       "      <td>108.15</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7795-CFOCW</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>45</td>\n",
       "      <td>No</td>\n",
       "      <td>No phone service</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Bank transfer (automatic)</td>\n",
       "      <td>42.30</td>\n",
       "      <td>1840.75</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>2</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Fiber optic</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>70.70</td>\n",
       "      <td>151.65</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   customerID  gender  SeniorCitizen Partner Dependents  tenure PhoneService  \\\n",
       "0  7590-VHVEG  Female              0     Yes         No       1           No   \n",
       "1  5575-GNVDE    Male              0      No         No      34          Yes   \n",
       "2  3668-QPYBK    Male              0      No         No       2          Yes   \n",
       "3  7795-CFOCW    Male              0      No         No      45           No   \n",
       "4  9237-HQITU  Female              0      No         No       2          Yes   \n",
       "\n",
       "      MultipleLines InternetService OnlineSecurity  ... DeviceProtection  \\\n",
       "0  No phone service             DSL             No  ...               No   \n",
       "1                No             DSL            Yes  ...              Yes   \n",
       "2                No             DSL            Yes  ...               No   \n",
       "3  No phone service             DSL            Yes  ...              Yes   \n",
       "4                No     Fiber optic             No  ...               No   \n",
       "\n",
       "  TechSupport StreamingTV StreamingMovies        Contract PaperlessBilling  \\\n",
       "0          No          No              No  Month-to-month              Yes   \n",
       "1          No          No              No        One year               No   \n",
       "2          No          No              No  Month-to-month              Yes   \n",
       "3         Yes          No              No        One year               No   \n",
       "4          No          No              No  Month-to-month              Yes   \n",
       "\n",
       "               PaymentMethod MonthlyCharges  TotalCharges Churn  \n",
       "0           Electronic check          29.85         29.85    No  \n",
       "1               Mailed check          56.95        1889.5    No  \n",
       "2               Mailed check          53.85        108.15   Yes  \n",
       "3  Bank transfer (automatic)          42.30       1840.75    No  \n",
       "4           Electronic check          70.70        151.65   Yes  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('./data/TelcoCustomerChurn.csv')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before we go any further, I will share a description of each column as described on the dataset's [Kaggle](https://www.kaggle.com/blastchar/telco-customer-churn) page. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**customerID** Customer ID\n",
    "\n",
    "**gender** Whether the customer is a male or a female\n",
    "\n",
    "**SeniorCitizen** Whether the customer is a senior citizen or not (1, 0)\n",
    "\n",
    "**Partner** Whether the customer has a partner or not (Yes, No)\n",
    "\n",
    "**Dependents** Whether the customer has dependents or not (Yes, No)\n",
    "\n",
    "**tenure** Number of months the customer has stayed with the company\n",
    "\n",
    "**PhoneService** Whether the customer has a phone service or not (Yes, No)\n",
    "\n",
    "**MultipleLines** Whether the customer has multiple lines or not (Yes, No, No phone service)\n",
    "\n",
    "**InternetService** Customer’s internet service provider (DSL, Fiber optic, No)\n",
    "\n",
    "**OnlineSecurity** Whether the customer has online security or not (Yes, No, No internet service)\n",
    "\n",
    "**OnlineBackup** Whether the customer has online backup or not (Yes, No, No internet service)\n",
    "\n",
    "**DeviceProtection** Whether the customer has device protection or not (Yes, No, No internet service)\n",
    "\n",
    "**TechSupport** Whether the customer has tech support or not (Yes, No, No internet service)\n",
    "\n",
    "**StreamingTV** Whether the customer has streaming TV or not (Yes, No, No internet service)\n",
    "\n",
    "**StreamingMovies** Whether the customer has streaming movies or not (Yes, No, No internet service)\n",
    "\n",
    "**Contract** The contract term of the customer (Month-to-month, One year, Two year)\n",
    "\n",
    "**PaperlessBilling** Whether the customer has paperless billing or not (Yes, No)\n",
    "\n",
    "**PaymentMethod** The customer’s payment method (Electronic check, Mailed check, Bank transfer (automatic), Credit card (automatic))\n",
    "\n",
    "**MonthlyCharges** The amount charged to the customer monthly\n",
    "\n",
    "**TotalCharges** The total amount charged to the customer\n",
    "\n",
    "**Churn** Whether the customer churned or not (Yes or No)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### First, let's view the size of this dataset, find out if anything is missing from it, and if we have any duplicate entries."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "We have 7043 rows (observations) of data with 21 columns (features).\n",
      "\n",
      "Number of duplicate rows: 0\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print('We have %d rows (observations) of data with %d columns (features).\\n' % (data.shape[0], data.shape[1]))\n",
    "dups = data[data.duplicated()]\n",
    "print(\"Number of duplicate rows: %d\\n\" % len(dups))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 7043 entries, 0 to 7042\n",
      "Data columns (total 21 columns):\n",
      " #   Column            Non-Null Count  Dtype  \n",
      "---  ------            --------------  -----  \n",
      " 0   customerID        7043 non-null   object \n",
      " 1   gender            7043 non-null   object \n",
      " 2   SeniorCitizen     7043 non-null   int64  \n",
      " 3   Partner           7043 non-null   object \n",
      " 4   Dependents        7043 non-null   object \n",
      " 5   tenure            7043 non-null   int64  \n",
      " 6   PhoneService      7043 non-null   object \n",
      " 7   MultipleLines     7043 non-null   object \n",
      " 8   InternetService   7043 non-null   object \n",
      " 9   OnlineSecurity    7043 non-null   object \n",
      " 10  OnlineBackup      7043 non-null   object \n",
      " 11  DeviceProtection  7043 non-null   object \n",
      " 12  TechSupport       7043 non-null   object \n",
      " 13  StreamingTV       7043 non-null   object \n",
      " 14  StreamingMovies   7043 non-null   object \n",
      " 15  Contract          7043 non-null   object \n",
      " 16  PaperlessBilling  7043 non-null   object \n",
      " 17  PaymentMethod     7043 non-null   object \n",
      " 18  MonthlyCharges    7043 non-null   float64\n",
      " 19  TotalCharges      7043 non-null   object \n",
      " 20  Churn             7043 non-null   object \n",
      "dtypes: float64(1), int64(2), object(18)\n",
      "memory usage: 1.1+ MB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SeniorCitizen</th>\n",
       "      <th>tenure</th>\n",
       "      <th>MonthlyCharges</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.162147</td>\n",
       "      <td>32.371149</td>\n",
       "      <td>64.761692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.368612</td>\n",
       "      <td>24.559481</td>\n",
       "      <td>30.090047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>18.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>35.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>29.000000</td>\n",
       "      <td>70.350000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>55.000000</td>\n",
       "      <td>89.850000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>72.000000</td>\n",
       "      <td>118.750000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       SeniorCitizen       tenure  MonthlyCharges\n",
       "count    7043.000000  7043.000000     7043.000000\n",
       "mean        0.162147    32.371149       64.761692\n",
       "std         0.368612    24.559481       30.090047\n",
       "min         0.000000     0.000000       18.250000\n",
       "25%         0.000000     9.000000       35.500000\n",
       "50%         0.000000    29.000000       70.350000\n",
       "75%         0.000000    55.000000       89.850000\n",
       "max         1.000000    72.000000      118.750000"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Also, we want to get a look at the target variable before we go any further."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "No     5174\n",
       "Yes    1869\n",
       "Name: Churn, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['Churn'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, right from the start, we see we have a major imbalance in the dataset. Let's visualize this."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1,2)\n",
    "fig.set_figheight(6)\n",
    "fig.set_figwidth(12)\n",
    "\n",
    "ax1.pie(data['Churn'].value_counts(), explode=(0,0.1), autopct='%1.1f%%',\n",
    "        shadow=True, startangle=90, labels=data['Churn'].unique(), colors=col_scheme)\n",
    "fig.set_facecolor('white')\n",
    "\n",
    "ax2.bar(data['Churn'].unique(), data['Churn'].value_counts(), color=col_scheme)\n",
    "ax1.text(0.5, 0.0, \"Number of customers who stayed = %d\" % len(data[data['Churn'] == 'No']), size=12, ha=\"center\", \n",
    "         transform=ax1.transAxes)\n",
    "ax1.text(0.5,-0.1, \"Number of customers who left = %d\" % len(data[data['Churn'] == 'Yes']), size=12, ha=\"center\", \n",
    "         transform=ax1.transAxes)\n",
    "\n",
    "plt.suptitle('Churn')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we'll display some plots for the various columns to get a good understanding of how the data is distributed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (9,6))\n",
    "ten_mean = np.mean(data['tenure'])\n",
    "ten_med = np.median(data['tenure'])\n",
    "plt.hist(data[\"tenure\"], bins = 25, alpha = 0.65, color = 'green')\n",
    "plt.title(\"tenure\", fontsize = 15)\n",
    "plt.ylabel(\"Number of customers\", fontsize = 15)\n",
    "plt.xlabel(\"Months\", fontsize = 15)\n",
    "plt.xticks(np.arange(0, 75, step=6))\n",
    "plt.axvline(ten_mean, color='red', linestyle='solid', linewidth=2, label = \"Mean %.2f\" % ten_mean)\n",
    "plt.axvline(ten_med, color='blue', linestyle='solid', linewidth=2, label = \"Median %.2f\" % ten_med)\n",
    "plt.legend(fontsize = 12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of customers who left before 6 months? : 19.47%\n"
     ]
    }
   ],
   "source": [
    "print(\"Percentage of customers who left before 6 months? : %.2f%%\" % ((len(data[data['tenure'] < 6])) / len(data['tenure']) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Wow, that is A LOT of customers leaving in their first 5 months with the company. It looks like something happens around the 6 year mark that is also causing people to leave, but we can't automatically jump to the conclusion in this case; it could be an effect of data collection or some other unknown cause. Let's focus on those people leaving in the first 5 months and see if they are breaking contract or leaving their Month-to-month plans."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "under_6 = data[data['tenure'] < 6]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_palette(sns.color_palette(col_scheme))\n",
    "plt.figure(figsize=(9,6))\n",
    "sns.countplot(x=under_6['Contract'], hue=under_6['Churn']).set_title('Churn vs. PaymentMethod')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The good news here is that there aren't that many customers who become so disgruntled that they decide to break contract. The amount of Month-to-month customers leaving before 6 months is still alarming though."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAm8AAAGKCAYAAABaV6YkAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAAA8d0lEQVR4nO3de5xVZb348c9XQEUhEFAkESUv5S1TKfF4MlKPtyxLzUxTsIua2Umr4zV/2sXK1ExPx8wSBRPJu1akYmiZHU0svKIHyht4AUVUAm/w/f2x1oybYWbYMHvPzMbP+/Xar73Xs9Zez/eZvWf48jzPWk9kJpIkSWoMq3V1AJIkSaqeyZskSVIDMXmTJElqICZvkiRJDcTkTZIkqYGYvEmSJDUQkzdJNRcRZ0RERsSMNvbPKPefUYe6j4yIT7ZS/kREnLOc944q49p6JeodHBE/iYh/RMTrEfFSRNwaEQdWHHNZRExd0XNLUiWTN0n18howPCJGVBZGxAeBjcv99XAk8Mk6nbtVEfFe4O/Ax4BzgD2Aw4F/AFdExLadGY+kVVvPrg5A0irrX8DfgIOByt6mg4EpwA5dEVSdXAHMA/4tM1+pKP9NRPwMmF/vACKiF7AkMxfXuy5JXcueN0n1NBE4KCICoHw+qCxfSkQcFBEPlkOOT0fEmRHRs2L/mHJIc5uImBwR/4qIRyNi/4pj7qBICkeXx2ZEjGlRz/ERMasc1pwYEf3bCj4irirP2bL8jIh4PiJ6RcQuZZ0nt0jcAMjMBzLzqRbv/4+IeKBsw58jYqsW+78REfdGxMtlPb+JiE1bHHNHRFxTDhP/g6In891R+G5EzImIVyJibEQcXP4sNq54/5oR8aPyZ/16RNwfEfu0qOMTEXFfGedLEXFPRHykrZ+XpM5h8iapnq4DBgP/Xm5/GFi3LG8WEXsAv6boqdsP+G/gm8BPWznnBOAm4FPADGBiRAwt9x0DPApMAnYqH7+reO9BwG4UQ6snAvsC328n/kuAXSJieEWsAYwGfpWZbwIfARYDt7VznkrDgLOBM4HPAusBv25KcEtDKdq+H/AloAfwl4jo1+JcOwNfLtvyceBl4DjgFOAi4EBgEfCjVuK4BhhD0f6PA/cCN0XEB8p2blIeM6XcfyjwW2BAle2UVCcOm0qqm8ycHxE3UwyV3lk+35yZLy+dq/Ad4I7MHF1u31zu/0FEfC8zZ1Uce15mjgWIiPuA5ymSsIsy85GI+BcwNzPvbiWkN4FPZuZb5fu3LGM6po0mTAZmUSQ5p5dlH6WYs3dpub1BWd+i5f08SgOAnTNzRhnDasD1wHspEk8y8/imgyOiRxnHHIpkbnzFufoDH8jM5yuOPYHiZ/H/ymNuLZPPDSvOuRvF/LxRmfnHiuM2B04FPg1sB7yamf9VUd+kKtsoqY7seZNUbxOBAyNiDYqeoKWGTMuEY3vg6hbv+zXF36idWpTf2vQiM1+kSGqGUp3bmxK30iPAeuV8sWVk5hKKJO3wip6xMcDUzHyo8tAq6wd4oilxq4gBKtoQESPLoeEXgbeAhUAfYPMW57qvKXErbQisT9EzWanl9u7Ac8BdEdGz6QH8AWi6wORBoF9EjIuIPSJi7RVoo6Q6MnmTVG83USQeZwJrA79psX8Q0IuiB61S03bLYbr5LbbfANasMpbW3hvAGu2851JgI+CjEdEXOAAYW7F/NrBuRHQkBijbEBHDKBLUAI6iGBr9IEWS2rKOlj+z9cvnuS3KW24PKo99s8XjDMoeusx8jKKn7z0UPW4vRMSEiFh3Oe2TVGcOm0qqq8z8V0T8FjgeuDoz/9XikBcoEof1WpQPLp/n1TnEdmXmExFxG0WP23CK//ReWXHIHRTDvrux9Py6lbUXsBawX9PPquwVa22uWcsev+fK55YJVsvteRRJ5yfbCyQzfwf8rpxr9zHgJxTzEQ9utwWS6sqeN0md4WcUPW4XtdxR3triPop5VpUOApYA/7uCda1IT1y1LqHocTsGuCEz5zftyMw7KeL/ftkzt5Ty6tgNW5a3ozdFuyuHdw+iuv9sP02RwO3XovwTLbb/QNHztiAzp7Z8tDxpZr6cmRMo5uZtWWU7JNWJPW+S6i4z76DooWrL6cAtEXEpxZy4bYDvAr9ocbFCNR4F9oyIPYEXgcfLuXEdcQNwIcXcvJNb2X8ocDswNSLOo5jH9i5gT4qrRXekSKyqMYXi6tJLI+ISYCuKK2/nL++Nmbk4Is4Gzo6IucBdFInbNuUhS8rnycAtwOSIOAt4uIz3A8CamXlyRBxFMd/wZuAZYDOKBLvygglJXcCeN0ldLjNvpRiKG0HRQ3cccC5w7Eqc7nvAdOAqittffLwG8b0O/J4iAVvmliDl/LDtKRKdEyiSo8spLjA4JDPvX4G6HqQYot2R4tYch1AkTS9XeYrzgB9Q9BJeC6zD27dDeaWsI4H9KebuHUeRyP2cIln7c3nsAxTDrT+mmIP3LeAXFLclkdSFovgdliS1pZxz9iQwNjNP6+p4VlRE/BL4j8zcqKtjkdRxDptKUhsiYnVgW4rer4EUvVPdWkRsDXwG+AvFMOnewBHYYyatMkzeJKlt7wb+SnGbjqNWYv5dV/gXxYoWx1LcmuVJisTt3K4MSlLtOGwqSZLUQLxgQZIkqYGYvEmSJDWQd8yct0GDBuXGG2/c1WFI6g7uu6943mGHVbE6SauA++6774XMbHU5unfMnLcRI0bk1KnL3Dhc0jtR0xrznfT3r5Ork7QKiIj7MnNEa/scNpUkSWogJm+SJEkNxORNkiSpgZi8SZIkNZB3zNWmkiQ1gjfffJNZs2bx2muvdXUoqrM111yToUOH0qtXrxV6n8mbJEndyKxZs+jbty8bb7wx0XSpslY5mcmLL77IrFmzGD58+Aq9t1OHTSNibETMiYiHKsrOjohHI+KBiLg+IvpX7Ds5ImZGxGMRsWdF+V5l2cyIOKkz2yBJUj299tprDBw40MRtFRcRDBw4cKV6WDt7zttlwF4tyiYDW2fm+4H/A04GiIgtgYOBrcr3XBgRPSKiB/A/wN7AlsBny2MlSVolmLi9M6zs59ypyVtm/gmY16Ls1sx8q9y8Gxhavt4PmJiZr2fm48BM4EPlY2Zm/jMz3wAmlsdKkiSt8rrb1aafB35fvt4AeLpi36yyrK3yZUTEkRExNSKmzp07tw7hSpIkda5uk7xFxKnAW8AVtTpnZl6cmSMyc8S667a6PJgkSarSxhtvzOqrr84LL7ywVPl2221HRPDEE090ekwLFy7kmGOOYdCgQfTr149ddtllmWPeeOMNtthiC4YOHdrKGQrf//736dOnT/Ojd+/erLbaaku19bbbbmP77bdn7bXXZujQoVx11VV1adPydIvkLSLGAPsCh+bbi63OBjasOGxoWdZWuSRJqrPhw4dz5ZVXNm8/+OCDLFy4sMviOfLII5k3bx7Tp09n3rx5nHfeecscc/bZZ7O8TpxTTjmFBQsWND9OPPFERo0axaBBgwB45JFHOOSQQzjzzDN5+eWXuf/++9lhhx3q0qbl6fLkLSL2Ak4APpGZlZ/+TcDBEbFGRAwHNgP+CtwLbBYRwyNidYqLGm7q7LglSXonOuywwxg/fnzz9rhx4zj88MOXOub111/nm9/8JsOGDWPw4MEcffTRLFq0CICXXnqJfffdl3XXXZd11lmHfffdl1mzZjW/d9SoUZx22mnsvPPO9O3blz322GOZnr4mjz76KDfddBMXX3wx6667Lj169FgmoXr88cf51a9+xcknn1x1GzOT8ePHM3r06Oay733vexx11FHsvffe9OzZk4EDB7LJJptUfc5a6tT7vEXElcAoYFBEzAJOp7i6dA1gcnnVxd2ZeXRmPhwRVwGPUAynfiUzF5fnORa4BegBjM3MhzuzHe055NpDanq+CQdMqOn5JEkNpLOuOm0e9Fq+kSNHcvnllzN9+nQ233xzJk6cyF133cW3vvWt5mNOOukk/vGPfzBt2jR69erFIYccwne+8x1+8IMfsGTJEo444giuuuoqFi9ezOc//3mOPfZYbrjhhub3T5gwgd///vdsuOGG7L333pxzzjn88Ic/XCaWv/71r2y00UacfvrpXH755QwZMoQzzjiDAw44oPmYr371q3z/+9+nd+/eVbfxzjvvZM6cOUud5+6772aTTTZhm2224YUXXmC33XbjggsuYMCAAVWft1Y6+2rTz2bmkMzslZlDM/OSzNw0MzfMzA+Uj6Mrjj8zMzfJzPdm5u8ryidl5ublvjM7sw2SJL3TNfW+TZ48mS222IINNnj7usHM5OKLL+a8885jwIAB9O3bl1NOOYWJEycCMHDgQA444ADWWmst+vbty6mnnsof//jHpc5/xBFHsPnmm9O7d28OOuggpk2b1mocs2bN4qGHHqJfv34888wz/PSnP2X06NFMnz4dgOuvv57FixfzqU99aoXaN27cOA488ED69OmzVF2XX3451157LTNmzGDRokV89atfXaHz1oorLEiS1F2tQI9YZzrssMPYZZddePzxx5cZMp07dy4LFy5cavgyM1m8eDFQXGBw/PHHc/PNN/PSSy8B8Oqrr7J48WJ69OgBwPrrr9/83rXWWosFCxa0Gkfv3r3p1asX3/rWt+jZsycf+chH+OhHP8qtt97KsGHDOOGEE5g0adIKtW3hwoVcffXV3HjjjcvU1ZRUQjFHbvfdd1+hc9eKyZskSVohG220EcOHD2fSpElccsklS+0bNGgQvXv35uGHH16qR67Jueeey2OPPcY999zD+uuvz7Rp09huu+3IlUhU3//+9y9T1nTj2xkzZvDEE0/w4Q9/GCiuOH355ZdZf/31ufvuu9l4441bPef111/PgAEDGDVq1DJ1Vd5UtytvpNzlFyxIkqTGc8kllzBlyhTWXnvtpcpXW201vvSlL3H88cczZ84cAGbPns0tt9wCFL1svXv3pn///sybN49vf/vbKx3DLrvswrBhw/jBD37AW2+9xV133cXtt9/OnnvuydZbb83TTz/NtGnTmDZtGr/85S8ZPHgw06ZNY8MNN2zznE0XYLRMzo444gguvfRS/vnPf7Jw4UJ++MMfsu+++6507B1h8iZJklbYJptswogRI1rdd9ZZZ7HpppsycuRI3vWud7H77rvz2GOPAXDcccexaNEiBg0axMiRI9lrr5arZlavV69e3HjjjUyaNIl+/frxpS99ifHjx/O+972Pnj17sv766zc/BgwYwGqrrcb666/fPDzbp08f7rzzzubzzZ49mylTpiwzFAzw+c9/nsMPP5wdd9yRjTbaiDXWWIMLLrhgpWPviFiZbspGNGLEiJw6dWrd6/FqU6kBNP2PupP+/nVydWpw06dPZ4sttujqMNRJ2vq8I+K+zGw1O7bnTZIkqYGYvEmSJDUQkzdJkqQGYvImSZLUQEzeJEmSGojJmyRJUgMxeZMkSWogJm+SJEkNxORNkiR1qYhg5syZABx99NF897vf7eKIujeTN0mSVJWNN96Y1VdfnRdeeGGp8u22246I4IknnuhwHRdddBGnnXZah8/T0hVXXEGfPn2aH2uttRYRwX333QdAZnLiiScycOBABg4cyIknnkhbq1DdfvvtbLPNNvTv35+BAwfyqU99itmzZzfvHzNmDKuvvvpS9S1evLhmbTF5kyRJVRs+fDhXXnll8/aDDz7IwoULuzCi6hx66KEsWLCg+XHhhRfynve8h+233x6Aiy++mBtuuIH777+fBx54gN/85jf8/Oc/b/VcW265Jbfccgvz58/nmWeeYbPNNuPLX/7yUseccMIJS9XXtJ5qLZi8SZKkqh122GGMHz++eXvcuHHLLOT++uuv881vfpNhw4YxePBgjj76aBYtWtS8/+yzz2bIkCG8+93vZuzYsUu9d8yYMXzrW98C4KWXXmLfffdl3XXXZZ111mHfffdl1qxZzceOGjWK0047jZ133pm+ffuyxx57LNMr2JamuKNcfHjcuHF84xvfYOjQoWywwQZ84xvf4LLLLmv1vYMHD+bd735383aPHj2ah307g8mbJEndVETnPFbEyJEjeeWVV5g+fTqLFy9m4sSJfO5zn1vqmJNOOon/+7//Y9q0acycOZPZs2fzne98B4Cbb76Zc845h8mTJzNjxgxuu+22NutasmQJRxxxBE8++SRPPfUUvXv35thjj13qmAkTJnDppZcyZ84c3njjDc4555zltuHJJ5/kT3/601JJ58MPP8y2227bvL3tttvy8MMPt3mOp556iv79+9O7d2/OOeccTjjhhKX2X3jhhQwYMIAddtiBa6+9drkxrQiTN0mStEKaet8mT57MFltswQYbbNC8LzO5+OKLOe+88xgwYAB9+/bllFNOYeLEiQBcddVVHHHEEWy99dasvfbanHHGGW3WM3DgQA444ADWWmst+vbty6mnnsof//jHpY454ogj2HzzzenduzcHHXQQ06ZNW27848eP58Mf/jDDhw9vLluwYAH9+vVr3u7Xrx8LFixoc97bsGHDmD9/Pi+88ALf+973eN/73te87z//8z+ZMWMGc+bM4bvf/S5jxozhrrvuWm5c1epZszNJkqSaaiNv6HKHHXYYu+yyC48//vgyQ6Zz585l4cKF7LDDDs1lmdk8Yf+ZZ55Zat9GG23UZj0LFy7k+OOP5+abb+all14C4NVXX2Xx4sXNc8jWX3/95uPXWmstFixYsNz4x48fzymnnLJUWZ8+fXjllVeat1955RX69OnTPKzalgEDBjB69Gi23XZbZs+eTc+ePZvn0QHss88+HHrooVx33XXsvPPOy42tGva8SZKkFbLRRhsxfPhwJk2axP7777/UvkGDBtG7d28efvhh5s+fz/z583n55Zebk6ohQ4bw9NNPNx//1FNPtVnPueeey2OPPcY999zDK6+8wp/+9CeANnvDqnHXXXfxzDPPcOCBBy5VvtVWW3H//fc3b99///1stdVWVZ3zrbfeYs6cOUslf5UiokMxt2TyJkmSVtgll1zClClTWHvttZcqX2211fjSl77E8ccfz5w5cwCYPXs2t9xyCwAHHXQQl112GY888ggLFy7k29/+dpt1vPrqq/Tu3Zv+/fszb968do+t1rhx4zjggAPo27fvUuWHH344P/7xj5k9ezbPPPMM5557LmPGjGn1HNdddx2PPfYYS5YsYe7cuXz9619nu+22Y8CAAQBcc801LFiwgCVLlnDrrbfyq1/9ik984hMdjr2JyZskSVphm2yyCSNGjGh131lnncWmm27KyJEjede73sXuu+/OY489BsDee+/Ncccdx6677sqmm27Krrvu2mYdxx13HIsWLWLQoEGMHDmSvfbaq0Mxv/baa1x11VWMHj16mX1HHXUUH//4x9lmm23Yeuut+djHPsZRRx3VvL9Pnz7ceeedQJGM7rXXXvTt25dtttmG1VZbjeuvv7752PPPP58NNtiA/v3781//9V/84he/YNSoUR2KvVLUshuvOxsxYkROnTq17vUccu0hNT3fhAMm1PR8knj78rpO+vvXydWpwU2fPp0tttiiq8NQJ2nr846I+zKz1ezYnjdJkqQGYvImSZLUQEzeJEmSGojJmyRJUgMxeZMkqZt5p1xM+E63sp+zyZskSd3ImmuuyYsvvmgCt4rLTF588UXWXHPNFX6vy2NJktSNDB06lFmzZjF37tyuDkV1tuaaazJ06NAVfp/JmyRJ3UivXr2WWjBdaslhU0mSpAZi8iZJktRATN4kSZIaiMmbJElSAzF5kyRJaiAmb5IkSQ3E5E2SJKmBmLxJkiQ1EJM3SZKkBmLyJkmS1EBM3iRJkhpIpyZvETE2IuZExEMVZQMiYnJEzCif1ynLIyIuiIiZEfFARGxf8Z7R5fEzImJ0Z7ZBkiSpK3V2z9tlwF4tyk4C/pCZmwF/KLcB9gY2Kx9HAj+DItkDTgd2BD4EnN6U8EmSJK3qOjV5y8w/AfNaFO8HjCtfjwM+WVE+Pgt3A/0jYgiwJzA5M+dl5kvAZJZNCCVJklZJ3WHO2+DMfLZ8/RwwuHy9AfB0xXGzyrK2ypcREUdGxNSImDp37tzaRi1JktQFukPy1iwzE8ganu/izByRmSPWXXfdWp1WkiSpy3SH5O35cjiU8nlOWT4b2LDiuKFlWVvlkiRJq7zukLzdBDRdMToauLGi/PDyqtORwMvl8OotwB4RsU55ocIeZZkkSdIqr2dnVhYRVwKjgEERMYviqtEfAldFxBeAJ4GDysMnAfsAM4GFwBEAmTkvIr4L3Fse953MbHkRhCRJ0iqpU5O3zPxsG7t2a+XYBL7SxnnGAmNrGJokSVJD6A7DppIkSaqSyZskSVIDMXmTJElqICZvkiRJDcTkTZIkqYGYvEmSJDUQkzdJkqQGYvImSZLUQEzeJEmSGojJmyRJUgMxeZMkSWogJm+SJEkNxORNkiSpgZi8SZIkNZCqkreIWC8ihldsR0QcGRE/iYiP1y88SZIkVaq25+0y4PiK7e8AFwJ7AddHxJjahiVJkqTWVJu8bQ9MAYiI1YCjgVMy833AmcBxdYlOkiRJS6k2eesHvFi+3gEYAFxRbk8BNq1xXJIkSWpFtcnbLGDL8vXHgEczc3a53Q94rdaBSZIkaVk9qzxuLHB2ROxOkbydXLFvJDC91oFJkiRpWVUlb5n5g4iYBXwI+CpFMtdkAPDLOsQmSZKkFpabvEXEGsA3gd9m5uUt92fm0fUITJIkScta7py3zHwdOBXoX/doJEmS1K5qL1i4h+J2IZIkSepC1V6wcAIwISLeBCYBzwNZeUBmLqxxbJIkSWqh2uTtnvL5AuD8No7p0fFwJEmS1J5qk7fP06KnTZIkSZ2v2luFXFbnOCRJklSFanveAIiILSmWx9oQGJuZz0XEpsDzmflqPQKUJEnS26pK3iKiD8WNeQ8E3izfdzPwHPB94CmKe8FJkiSpjqq9VciPgX8DdgP6AlGxbxKwV43jkiRJUiuqHTbdH/haZt4eES2vKn0S2Ki2YUmSJKk11fa89QZebGNfX2BxbcKRJElSe6pN3u4FDm9j34HAX2oTjiRJktpT7bDpacDkiLgNuJrinm/7RMTxFMnbLnWKT5IkSRWq6nnLzDspLlZYA/gpxQUL3wbeA+yemffWLUJJkiQ1q/o+b5l5F/DhiOgNrAPMdz1TSZKkzrVCN+kFyMxFwKI6xCJJkqTlqDp5i4gPUtwyZANgzRa7MzM/U8vAJEmStKxqV1g4HjgXeB74J/BGPYOSJElS66rtefsGcD7w9czMOsYjSZKkdlR7n7c1gN+ZuEmSJHWtapO3yyjmu0mSJKkLVTtseiLw0/ImvVOA+S32Z2b+rCOBlPPqvkhxA+AHgSOAIcBEYCBwH3BYZr4REWsA44EdKJbt+kxmPtGR+iVJkhpBtcnbrsChFOuY7trK/gRWOnmLiA2A/wS2zMxFEXEVcDCwD3BeZk6MiIuAL5T1fAF4KTM3jYiDgbMAr3aVJEmrvGqHTS8E7gG2AtbIzNVaPHrUIJaeQO+I6AmsBTxLkSheU+4fB3yyfL1fuU25f7eIiBrEIEmS1K1Vm7y9G/hRZk7PzDdrHURmzgbOAZ6iSNpephgmnZ+Zb5WHzaK4xxzl89Ple98qjx9Y67gkSZK6m2qTt9uAbesVRESsQ9GbNpwiUVwb2KsG5z0yIqZGxNS5c+d29HSSJEldrto5bxcAF5XrmrZ2wQKZ+UgH4tgdeDwz5wJExHXAzkD/iOhZ9q4NBWaXx88GNgRmlcOs/SguXGgZ08XAxQAjRozwNieSJKnhVZu83VY+fwf4dot9QXHBQkfmvT0FjIyItSjWTd0NmArcDhxIccXpaODG8vibyu3/LfdP8R50kiTpnaDa5O2j9QwiM++JiGuAvwFvAX+n6DH7HTAxIr5Xll1SvuUS4PKImAnMo7gyVZIkaZVXVfKWmX+sdyCZeTpweovifwIfauXY14BP1zsmSZKk7qbanjcAImJH4N+BARQ9Xn/OzHvqEZgkSZKWVVXyFhFrA1dTXAH6FsXFAQOBHhFxM/DpzFxYtyglSZIEVH+rkB8BO1GsYrBmZg4B1qSYa7YTxQoHkiRJqrNqk7cDgBMz8+rMXAKQmUsy82rgJJx/JkmS1CmqTd76Ua5o0IqngXfVJhxJkiS1p9rk7X7gyy3XDy23v1zulyRJUp1Ve7XpKcDvgUcj4nrgeWA94FPAxsDedYlOkiRJS6n2Pm9TImJ74DSK+W1DKBaQvwfYv4NLY0mSJKlKVd/nLTMfxpUMJEmSulRVc94iYkpEvK+NfZtHxJTahiVJkqTWVHvBwijavqL0XcAuNYlGkiRJ7ao2eQPIlgURsTqwK/BczSKSJElSm9qc8xYRpwP/r9xM4O4WdwqpdHaN45IkSVIr2rtgYRLwAhDABcC5wBMtjnkDeDQz76xLdJIkSVpKm8lbZt4L3AsQEa8Cv8vMFzorMEmSJC2r2luFXAH0qCyIiD2ALYE/Zebfah2YJEmSllVt8vZr4GXg8wAR8Z/AT4DXgR4RsX9m/rYuEUqSJKlZtVebjqSYA9fkv4BzM7M38Evg1FoHJkmSpGVVm7wNpLwdSERsA7wbuKjcdzXF8KkkSZLqrNrk7XmKBegB9gKezMx/lNu9gSU1jkuSJEmtqHbO29XAWRGxLXAE8NOKfdsBM2odmCRJkpZVbfJ2EvAK8EHgZ8APKvbtQHFBgyRJkuqsquQtM98CvtPGvv1rGpEkSZLaVFXyFhFrLe+YzFzY8XAkSZLUnmqHTRfQysL0LfRYzn5JkiR1ULXJ2+dZNnlbB9iT4jYh361lUJIkSWpdtXPeLmtj108i4mfAVjWLSJIkSW2q9j5v7bkWOLwG55EkSdJy1CJ5+yDFGqeSJEmqs2qvNv1RK8WrA1sAu1EsUi9JkqQ6q/aChU+3UvYaMAv4T+DimkUkSZKkNlV7wcLwegciSZKk5avFnDdJkiR1kqqSt4gYGxET29h3ZUT8orZhSZIkqTXV9rz9B8UtQVpzLcXNeiVJklRn1SZv6wLz2tj3ErBebcKRJElSe6pN3p4Edmlj3y4UV51KkiSpzqpN3i4DToyIr0REH4CI6BMRxwAnAL+sU3ySJEmqUO193s4CNgH+G7ggIv4FrA0ExT3ezqpPeJIkSapU7X3elgBfjIizgY8CA4EXgSmZ+X91jE+SJEkVqu15AyAzHwMeq1MskiRJWg5v0itJktRATN4kSZIaiMmbJElSA2kzeYuIYRHRq7MCiYj+EXFNRDwaEdMjYqeIGBARkyNiRvm8TnlsRMQFETEzIh6IiO07K05JkqSu1F7P2+PAdgARMSUi3lfnWM4Hbs7M9wHbAtOBk4A/ZOZmwB/KbYC9gc3Kx5HAz+ocmyRJUrfQXvK2CFirfD0KeFe9goiIfhQrNVwCkJlvZOZ8YD9gXHnYOOCT5ev9gPFZuBvoHxFD6hWfJElSd9HerUL+DpwfEZPL7a9GxLNtHJuZeWIH4hgOzAUujYhtgfuArwGDM7OpzueAweXrDYCnK94/qyxbKr6IOJKiZ45hw4Z1IDxJkqTuob3k7UvA2RS9XAnsBrzexrEJdCR56wlsD3w1M++JiPN5e4i0qCAzIyJX5KSZeTHFChCMGDFihd4rSZLUHbWZvGXmo8DHASJiCfDJzPxrneKYBczKzHvK7WsokrfnI2JIZj5bDovOKffPBjaseP/QskySJGmVVu2tQoYD0+oVRGY+BzwdEe8ti3YDHgFuAkaXZaOBG8vXNwGHl1edjgRerhhelSRJWmVVu7bpkxHRMyI+A/w7MACYB9wJXJeZb9Uglq8CV0TE6sA/gSMoksurIuILwJPAQeWxk4B9gJnAwvJYSZKkVV5VyVtErAfcCrwfeAJ4HtgJ+Apwf0TskZlzOxJIZk4DRrSya7dWjs2ybkmSpHeUaodNfwwMBEZm5nsyc6fMfA+wY1n+43oFKEmSpLdVm7ztA5zY8oKFzLwXOBn4WK0DkyRJ0rKqTd7WAF5tY9+rwOq1CUeSJEntqTZ5uxs4MSLWriwst08s90uSJKnOqrpgAfgGcDvF7TxupbhgYT1gTyAols+SJElSnVXV81ZeCboZxWoF6wL/QZG8XQRslpn31ytASZIkva3anjcy8wVaLFklSZKkzlXtnDdJkiR1AyZvkiRJDcTkTZIkqYGYvEmSJDWQ5SZvEbFGRJwaEdt2RkCSJElq23KTt8x8HTgV6F/3aCRJktSuaodN7wG2r2cgkiRJWr5q7/N2AjAhIt4EJlGssJCVB2TmwhrHJkmSpBaqTd7uKZ8vAM5v45geHQ9HkiRJ7ak2efs8LXraJEmS1PmqSt4y87I6xyFJkqQqVL22KUBEbAnsAGwIjM3M5yJiU+D5zHy1HgFKkiTpbVUlbxHRBxgLHAi8Wb7vZuA54PvAU8A36xSjJEmSStXeKuTHwL8BuwF9gajYNwnYq8ZxSZIkqRXVDpvuD3wtM2+PiJZXlT4JbFTbsCRJktSaanveegMvtrGvL7C4NuFIkiSpPdUmb/cCh7ex70DgL7UJR5IkSe2pdtj0NGByRNwGXE1xz7d9IuJ4iuRtlzrFJ0mSpApV9bxl5p0UFyusAfyU4oKFbwPvAXbPzHvrFqEkSZKaVX2ft8y8C/hwRPQG1gHmu56pJElS56p2zlul1yju9baoxrFIkiRpOapO3iJin4j4C0Xy9hzwWkT8JSI+VrfoJEmStJSqkreIOAr4DbAA+Brw6fJ5AXBTuV+SJEl1Vu2ct1OAn2fmMS3KL4qIi4BTgZ/XNDJJUqc65NpDanq+CQdMqOn5JBWqHTYdCFzfxr5rgQG1CUeSJEntqTZ5ux34SBv7PgL8qTbhSJIkqT1tDptGxJYVmxcAv4yIgcANwBxgPeBTwN7AF+sYoyRJkkrtzXl7iGIlhSYBHFU+stxucjPQcsF6SZIk1Vh7ydtHOy0KSZIkVaXN5C0z/9iZgUiSJGn5ql4eq0lE9ARWb1nuUlmSJEn1V+1NevtFxIUR8SzFCguvtvKQJElSnVXb83YZxS1BfgHMBN6oV0CSJElqW7XJ227AUZl5ZT2DkSRJUvuqvUnvU4Bz2iRJkrpYtcnbCcC3ImJYPYORJElS+6oaNs3MSRGxOzAzIp4A5rdyzIc6GkxE9ACmArMzc9+IGA5MpFhb9T7gsMx8IyLWAMYDOwAvAp/JzCc6Wr8kSVJ3V1XyFhHnAMcB91LfCxa+BkwH3lVunwWcl5kTI+Ii4AvAz8rnlzJz04g4uDzuM3WKqUsdcu0hNT3fhAMm1PR8ktSWWv/9Av+GSVD9BQtfBE7NzB/UK5CIGAp8DDgT+HpEBLAr0PTbPw44gyJ52698DXAN8NOIiMysXM5LkiRplVPtnLeFFMOW9fQTirl1S8rtgcD8zHyr3J4FbFC+3gB4GqDc/3J5vCRJ0iqt2uTtfODIsjes5iJiX2BOZtY0QYyIIyNiakRMnTt3bi1PLUmS1CWqHTYdBOwIPBYRd7DsBQuZmSd2II6dgU9ExD7AmhRz3s4H+kdEz7J3bSgwuzx+NrAhMKtcrqsfxYULLYO6GLgYYMSIEQ6pSpKkhldt8nYg8BbQC/iPVvYnsNLJW2aeDJwMEBGjgG9m5qERcXVZ90RgNHBj+Zabyu3/LfdPcb6bJEl6J6j2ViHD6x1IG04EJkbE94C/A5eU5ZcAl0fETGAecHAXxSdJktSpqu156zSZeQdwR/n6n8Ay94/LzNeAT3dqYJIkSd1Atfd5O2Z5x2TmhR0PR5IkSe2ptuftp+3sa5prZvImSZJUZ1XdKiQzV2v5AAYAnwXuB7asZ5CSJEkqrPSct8ycD/w6IvoBPwdG1SgmSZIktaHam/S253FgRA3OI0mSpOXoUPIWEUOAb1AkcJIkSaqzaq82ncvbFyY0WR3oC7wG7F/juCRJktSKaue8/Q/LJm+vUSwWf3NmLrM0lSRJtXbItYfU9HwTDphQ0/NJnaHaFRbOqHMckiSpFSasaqkWFyxIkiSpk7TZ8xYRU1bgPJmZu9UgHkmSJLWjvWHTauaxDQH+jWXnw0mSJKkO2kzeMrPNhd8jYhhwIrAv8AJwXu1DkyRJUksrtMJCRGwKnAx8DphTvv55Zi6qQ2ySpHbUeiK7pMZQ7X3etgJOBT4NPA18DRibmW/UMTZJ6jK1TYy8uk/dh1evNr52rzaNiB0i4jrgAWB74IvAZpl5kYmbJElS52vvatPfA3sADwIHZ+bVnRaVJEmSWtXesOme5fNQ4H8i4n/aO1FmrlezqCRJktSq9pK3b3daFJIkSapKe7cKMXmTJEnqZlboViGSJKlt3r5FncG1TSVJkhqIPW+SpHcse8rUiOx5kyRJaiAmb5IkSQ3E5E2SJKmBOOdNkjqJ86sk1YI9b5IkSQ3E5E2SJKmBmLxJkiQ1EJM3SZKkBmLyJkmS1EBM3iRJkhqIyZskSVID8T5v0kqo9f26JhwwoabnkyStukzeJC1XPW4ua8IqSSvHYVNJkqQGYvImSZLUQEzeJEmSGohz3iR1CS/6kKSVY8+bJElSAzF5kyRJaiAmb5IkSQ3E5E2SJKmBdIvkLSI2jIjbI+KRiHg4Ir5Wlg+IiMkRMaN8Xqcsj4i4ICJmRsQDEbF917ZAkiSpc3SXq03fAr6RmX+LiL7AfRExGRgD/CEzfxgRJwEnAScCewOblY8dgZ+Vz1JD8spLSVK1ukXPW2Y+m5l/K1+/CkwHNgD2A8aVh40DPlm+3g8Yn4W7gf4RMaRzo5YkSep83aXnrVlEbAxsB9wDDM7MZ8tdzwGDy9cbAE9XvG1WWfZsRRkRcSRwJMCwYcPqF7TUzdRjLVJJUvfQrZK3iOgDXAscl5mvRETzvszMiMgVOV9mXgxcDDBixIgVeq8kSVq+evxn0akf7es2yVtE9KJI3K7IzOvK4ucjYkhmPlsOi84py2cDG1a8fWhZJukdakX+AWn6Z8EeSkmNqFvMeYuii+0SYHpm/rhi103A6PL1aODGivLDy6tORwIvVwyvSpIkrbK6S8/bzsBhwIMRMa0sOwX4IXBVRHwBeBI4qNw3CdgHmAksBI7o1GglSZK6SLdI3jLzz0C0sXu3Vo5P4Ct1DUqSJKkb6hbDppIkSaqOyZskSVIDMXmTJElqICZvkiRJDcTkTZIkqYGYvEmSJDUQkzdJkqQGYvImSZLUQEzeJEmSGojJmyRJUgMxeZMkSWogJm+SJEkNxORNkiSpgZi8SZIkNRCTN0mSpAZi8iZJktRATN4kSZIaiMmbJElSA+nZ1QFIkiRVOuTaQ2p6vgkHTKjp+bqaPW+SJEkNxORNkiSpgZi8SZIkNRDnvGmVV+u5E5IkdSV73iRJkhqIyZskSVIDcdhU3Y7DnJIktc2eN0mSpAZi8iZJktRAHDZ9h6nHkOSqdudqSdKqZVVbscGeN0mSpAZi8iZJktRATN4kSZIaiMmbJElSAzF5kyRJaiBebaoO86a6kiR1HnveJEmSGojJmyRJUgMxeZMkSWogJm+SJEkNxORNkiSpgZi8SZIkNRCTN0mSpAZi8iZJktRAGjp5i4i9IuKxiJgZESd1dTySJEn11rDJW0T0AP4H2BvYEvhsRGzZtVFJkiTVV8Mmb8CHgJmZ+c/MfAOYCOzXxTFJkiTVVSMnbxsAT1dszyrLJEmSVlmRmV0dw0qJiAOBvTLzi+X2YcCOmXlsxTFHAkeWm+8FHuv0QFs3CHjBOrpNHZ1Vj3V0v3qso3vV0Vn1WEf3q8c6lrVRZq7b2o6eNaqgK8wGNqzYHlqWNcvMi4GLOzOoakTE1MwcYR3do47Oqsc6ul891tG96uiseqyj+9VjHSumkYdN7wU2i4jhEbE6cDBwUxfHJEmSVFcN2/OWmW9FxLHALUAPYGxmPtzFYUmSJNVVwyZvAJk5CZjU1XGshM4YyrWO7lePdXS/eqyje9XRWfVYR/erxzpWQMNesCBJkvRO1Mhz3iRJkt5xTN7qKCI2jIjbI+KRiHg4Ir5Wlg+IiMkRMaN8XqcOdZwREbMjYlr52KeDbVkzIv4aEfeX9Xy7LB8eEfeUS5T9urx4pNZ1XBYRj1e05QMdaUt5zh4R8feI+G2t29FOHfVoxxMR8WB5vqllWc2+X+3UUevvV/+IuCYiHo2I6RGxUx3a0VodtW7HeyvONS0iXomI42r8O99WHbVuy/Hl7+FDEXFl+ftZ09+TNuqo6e9JRHytPP/DEXFcWVbr71ZrdXT484iIsRExJyIeqihrNfYoXFB+Ng9ExPZ1qGNURLxc0ab/18G2fLr8mS2JiBEtjj+5bMtjEbFnreuIiI0jYlFFWy7qQB1nl39XHoiI6yOif0faUZXM9FGnBzAE2L583Rf4P4qlvH4EnFSWnwScVYc6zgC+WcO2BNCnfN0LuAcYCVwFHFyWXwR8uQ51XAYcWOPP5uvABOC35XbN2tFOHfVoxxPAoBZlNft+tVNHrb9f44Avlq9XB/rXoR2t1VHTdrSorwfwHLBRrdvSRh01awvFDc8fB3qX21cBY2r8+95WHTX7PQG2Bh4C1qKY430bsGktP4926ujw5wHsAmwPPFRR1mrswD7A7yn+jo4E7qlDHaMo/57VqC1bUNyD9Q5gREX5lsD9wBrAcOAfQI8a17Fx5XEdbMceQM/y9VkVP6+Vakc1D3ve6igzn83Mv5WvXwWmU/zB2o/iHxLK50/WoY6aysKCcrNX+UhgV+CasryjbWmrjpqKiKHAx4BflttBDdvRWh2drGbfr84QEf0o/iBeApCZb2TmfGrYjnbqqKfdgH9k5pPU7zOprKPWegK9I6InRWLyLDX+PWmljmc6eL6WtqBIYhZm5lvAH4H9qe3n0VYdHZaZfwLmtShuK/b9gPHl39G7gf4RMaTGday01urJzOmZ2drN8/cDJmbm65n5ODCTYknMWtaxUtqo49byswe4m+K+s7CS7aiGyVsniYiNge0oepMGZ+az5a7ngMF1qAPg2LIbd2xHhwXK8/eIiGnAHGAyxf8i5ld8aTu8RFnLOjKzqS1nlm05LyLW6EgdwE+AE4Al5fZAatyOVupoUst2QJHc3hoR90WxogjU/vvVWh1Qu+/XcGAucGkUw8y/jIi1qW072qoDavx7UuFg4MrydV1+51vUATVqS2bOBs4BnqJI2l4G7qOGvyet1ZGZt5a7a/V78hDw4YgYGBFrUfRObUhtP4+26oD6fLfair2WS0a29/PZKYqpLb+PiK1W8vzL01nLXw4v/x78MSI+XKNzfp6iBxTq2A6Tt04QEX2Aa4HjMvOVyn1Z9K12uHeplTp+BmwCfIDiD+O5Ha0jMxdn5gco/lfxIeB9HT3n8uqIiK2Bk8u6PggMAE5c2fNHxL7AnMy8rwbhrmgdNWtHhX/PzO2BvYGvRMQulTtr9P1qrY5afr96UgxD/CwztwP+RTFU06wG7Wirjpr/ngBEMRfsE8DVLffV8He+ZR01a0uZaOxHkfS+G1gb2KsD4VZVR0R8jhr+nmTmdIphrFuBm4FpwOIWx3To82injrp8t1rUXZPv0grU8TeKJZu2Bf4buKGeddfZs8Cw8u/B14EJEfGujpwwIk4F3gKuqEF87TJ5q7OI6EWRVF2RmdeVxc83dWeXz3NqXUdmPl8mQkuAX1Cjrtry3POB24GdKLrmm+4XuMwSZTWoY69yaDgz83XgUjrWlp2BT0TEE8BEimGg86ltO5apIyJ+VeN2AM29F2TmHOD68pw1/X61VkeNv1+zgFkVvazXUCRatWxHq3XU8fdkb+Bvmfl8uV3Tz6S1Omrclt2BxzNzbma+CVxH8b2u5e9Ja3X8W61/TzLzkszcITN3AV6imBdc69+RZeqo43errdiXu2RkR+vIzFeynNqSxX1We0XEoJWsoz21bEuryqHMF8vX91GMJG2+sueLiDHAvsChZcILdWyHyVsdlXOpLgGmZ+aPK3bdBIwuX48Gbqx1HS3mOnyKomt/pUXEuk1X0EREb+A/KObX3Q4cWB7W0ba0VsejFX9EgmLuxUq3JTNPzsyhmbkxxZDTlMw8lBq2o406PlfLdpTnWTsi+ja9ppg0+xC1/X61Wkctv1+Z+RzwdES8tyzaDXiEGrajrTpq/XtS4bMsPZxZs7a0VUeN2/IUMDIi1iq/r02fSc1+T9qoY3odfk/WK5+HUcxFm0CNP4/W6qjjd6ut2G8CDo/CSIph6GdbO8HK1hER65efCxHxIYoc4sWVrGN59R8cEWtExHBgM+Cvtayg/PemR/n6PWUd/1zJc+1FMU3mE5m5sGJX/dqRNbjqwUebV6X8O0V38wMUXenTKOZDDAT+AMyguDJpQB3quBx4sCy/CRjSwba8H/h7eb6HgP9Xlr+n/DLOpBi+WaMOdUwp2/IQ8CvKK1Jr8PmM4u0rQWvWjnbqqGk7ypjvLx8PA6eW5bX8frVVR62/Xx8AppbnuwFYp5btaKeOmrajrGdtin/Q+lWU1botrdVR68/k28Cj5ff1coor5mr6e9JGHbX+PbmTIvG8H9itTp9Ha3V0+POgSM6fBd6k6D3+QluxU1xl+j8UPUgPUnFlZQ3rOLb8O3A/xcT8f+tgWz5Vvn4deB64peL4U8u2PAbsXes6gAPKtkyjGA7+eAfqmEkxt21a+bioI+2o5uEKC5IkSQ3EYVNJkqQGYvImSZLUQEzeJEmSGojJmyRJUgMxeZMkSWogJm+SlhERZ0RERsSMNvbPKPefUaf6j4yIT7ZS/kREnLOc944qY9u6gzE0nafp8WZE/DMifhRvL63VdGxGxLEV23dExDUV22dExAsrUPcd5Tl370gbulpEbF62vX9XxyKtSkzeJLXlNYq1/0ZUFkbEB4GNy/31ciS1W7y9ow6lWE2kaTWOY4DzWhyzE60shbUyImIDoGmdxc/W4pxdaHPgdKB/F8chrVJM3iS15V8UN0w9uEX5wWX5vzo9oq7xQGbenZl3Zub5wIW0SCzL/c+3+u4V9xmKG65OAfYv1zCVpGYmb5LaMxE4qGJJnAAOKsuXEREHRcSDEfF6RDwdEWdWrIVJRIwphwO3iYjJEfGviHg0IvavOOYOYAdgdMWQ5ZgW9RwfEbMi4qWImNjesFxEXFWes2X5GRHxfBRrA6+IV4Gl3tNy2LSDPktx9/qzKHqs9m5RV9Nw7m4RcWP5M5wREXtERI+IODsiXoiI2RHx9ZYnr+IzanWIt5Wh4Sci4py2PouIGAX8pjz88fL9T3TwZyMJkzdJ7bsOGEyxDBsUw3nrluVLiYg9gF9TLDWzH/DfwDeBn7Zy3qb1JT9FsQTPxIgYWu47hmLJpEkUw5E7Ab+reO9BFOtgHgmcSLEY9PfbacMlwC7l2oJNsQbFmo2/ymJR9Pb0iIieEdE7InYGjgKuX857VkpEbAqMoEiO/0CxIHhbQ6c/B/5M8TN8EriG4mfdFzik3D43InasOP+KfEbVaO+z+Ft5bijW+9ypjFVSB/Vc/iGS3qkyc35E3EwxVHpn+XxzZr5cdsZV+g5wR2Y2LWh9c3nMDyLie5k5q+LY8zJzLEBE3Eex5uC+FGsCPhIR/wLmZubdrYT1JvDJzHyrfP+WZVzHtNGMyRRrEI6hmH8F8FGKeXuXLv+nwLQW238GvlbF+1bGZ4ElwFWZuTgirgaOiIi1M7PlMPXlmXk2QETMolin8b2ZuWtZdhvFEOz+wD3le1bkM6pGm59FZr4SEY+Vx/09M59YwXNLaoM9b5KWZyJwYESsARxIK0OmEdED2J5lJ+3/muLvzE4tym9tepGZL1L0MA2lOrc3JQulR4D12hr+zMwlFEna4fF2xjkGmJqZD1VR38HAB4EdKZKrQcANEVGPv5+fBf6Ymc+V2xOBtSh6yVr6Q8XrmeXzlKaCst3/BDaAlfqMqrFCn4Wk2jB5k7Q8NwF9gDOBtXl7HlOlQRTzwFpO2m/aHtCifH6L7TeANauMp7X3BrBGO++5FNgI+GhE9AUOAMZWWd/DmTk1M/+amRMprj7dFdinyvdXJSK2BbYAfhsR/cu5Yw8Dz9L60On8pheZ+UbLslLlz3VFP6NqtFbf8j4LSR1k8iapXeVw3W+B44HftDJ8B/ACxRDaei3KB5fP8+oX4fKVQ3a3UfS4HUTxt+/KlTzd9PJ5iw4HtrSmBO1c4KXyMQ8YAuwZESuTXFWq9jN6DVjqCteIWKeDdUuqIZM3SdX4GUWP20Wt7czMxcB9wKdb7DqIYg7X/65gfSvSE1etSyh63I4BbsjM+St5nqab/z5di6Cg+QKKg4HbKebjVT4OoegxO6AjdazAZzQL6Fveb67JHitZbVOPYK0/S+kdzQsWJC1XZt4B3LGcw04HbomISynmam0DfBf4xUpMhH+UordpT+BF4PFyblxH3EBxj7btgZNX4H3vj4g+FP/ZfQ9wGvAU8PsOxlNpJ4ph3RPLn/VSIuJkip65X3Swnmo+o5uBRcDYiDgXGA4cvZL1NV2wcFRETAQWZuaDKx29JMCeN0k1kpm3UvQejaDopTuOYghwZe5/9j2K4cmrgHuBj9cgvtcpEq6nKYZQq3UFRa/Un4EfUtwCY9fMfLmjMVX4LPAKxfzC1vwK+EhEDOlIJdV8Rpn5AkUv31CKhPdzFL1/K1PfkxS3C9kfuIvW50tKWkGRmV0dgyTVXXkj2ieBsZl5WlfHI0kry2FTSau0cnmpbSl6jwZS3NxWkhqWyZukVd27gb9S3EvuqJWYfydJ3YrDppIkSQ3ECxYkSZIaiMmbJElSAzF5kyRJaiAmb5IkSQ3E5E2SJKmBmLxJkiQ1kP8P0f+BwT290m0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (10,6))\n",
    "mc_mean = np.mean(data['MonthlyCharges'])\n",
    "mc_med = np.median(data['MonthlyCharges'])\n",
    "plt.hist(data[\"MonthlyCharges\"], bins = 25, alpha = 0.65, color = 'green')\n",
    "plt.xticks(np.arange(20, 125, step=5))\n",
    "plt.title(\"MonthlyCharges\", fontsize = 15)\n",
    "plt.ylabel(\"Number of customers\", fontsize = 15)\n",
    "plt.xlabel(\"Monthly Bill Amount\", fontsize = 15)\n",
    "plt.axvline(mc_mean, color='red', linestyle='solid', linewidth=2, label = \"Mean %.2f\" % mc_mean)\n",
    "plt.axvline(mc_med, color='blue', linestyle='solid', linewidth=2, label = \"Median %.2f\" % mc_med)\n",
    "plt.legend(fontsize = 12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of customers with a Monthly Bill of $25 or less? 19.96%\n"
     ]
    }
   ],
   "source": [
    "print(\"Percentage of customers with a Monthly Bill of $25 or less? %.2f%%\" % ((len(data[data['MonthlyCharges'] < 25.01])) / len(data['MonthlyCharges']) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For some odd reason, we run into an small issue here where **TotalCharges** is being read as a string variable. We'll coerce it to a float so we can proceed with our histogram."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert data in 'TotalCharges' from string to float.\n",
    "data['TotalCharges'] = pd.to_numeric(data['TotalCharges'],errors='coerce')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "How many missing values in 'TotalCharges' after coercion? : 11\n",
      "What percentage of the data are these missing values? : 0.16%\n"
     ]
    }
   ],
   "source": [
    "print(\"How many missing values in 'TotalCharges' after coercion? : %d\" % data['TotalCharges'].isnull().sum())\n",
    "print(\"What percentage of the data are these missing values? : %0.2f%%\" % ((data['TotalCharges'].isnull().sum()) / (len(data['TotalCharges'])) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After coercion of **TotalCharges** from string to float, we see that we actaully do have some missing values. It is safe to assume that these values are simply missing at random, so we’ll just remove these observations considering they only account for **0.16%** of the data.\n",
    "* Note: If we REALLY wanted to retain these rows, an easy fix is to fill the missing value with the mean or median. A more time consuming fix would be to construct a regression model that predicts these values - it would likely give us a better imputed value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "data2 = data[data[\"TotalCharges\"].notnull()] # Keep rows for which 'TotalCharges' has recorded data\n",
    "data2 = data2.reset_index(drop=True) # Reset the index of the dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (18,6))\n",
    "tc_mean = np.mean(data2['TotalCharges'])\n",
    "tc_med = np.median(data2['TotalCharges'])\n",
    "plt.hist(data2['TotalCharges'], bins = 25, alpha = 0.65, color = 'green')\n",
    "plt.title('TotalCharges', fontsize = 15)\n",
    "plt.ylabel(\"Number of customers\", fontsize = 15)\n",
    "plt.xlabel(\"Total Amount Charged\", fontsize = 15)\n",
    "plt.xticks(np.arange(0, 8800, 300))\n",
    "plt.axvline(tc_mean, color='red', linestyle='solid', linewidth=2, label = \"Mean %.2f\" % tc_mean)\n",
    "plt.axvline(tc_med, color='blue', linestyle='solid', linewidth=2, label = \"Median %.2f\" % tc_med)\n",
    "plt.legend(fontsize = 12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of customers with TotalCharges less than $300? 21.33%\n"
     ]
    }
   ],
   "source": [
    "print(\"Percentage of customers with TotalCharges less than $300? %.2f%%\" % ((len(data2[data2['TotalCharges'] < 300])) / len(data2['TotalCharges']) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We had to stretch out this histogram in order to get a better look, but we see that **21.33%** of the customers had **TotalCharges** less than **\\$300**."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With the visuals above we can gather that:\n",
    "* Getting these customers to stay for even just 6 months is a major problem. This accounts for **19.34%** of customers in the dataset.\n",
    "* The minimum amount a customer can pay per month is **\\\\$18.25** and **19.94%** of customers are paying no more than **\\\\$6.75** of that minimum amount. \n",
    "* About 1/5 of the customers in this data set have only been charged less than **\\\\$300** total, which makes sense when we consider the previous two points."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Something else we should look into are the columns that have a relationship with **InternetService**. For columns **OnlineSecurity**, **OnlineBackup**, **DeviceProtection**, **TechSupport**, **StreamingTV**, **StreamingMovies** the responses are: 'Yes', 'No', and 'No internet service'. We should make sure that if a customer has no internet service they also have a response for the above columns of 'No internet service'. If this is the case across the board, we can change these columns to binary variables with responses of simply 'Yes' or 'No'."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "no_www = data2[data2.InternetService == 'No'] # Subset customers that have no internet service"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['No internet service']\n",
      "['No internet service']\n",
      "['No internet service']\n",
      "['No internet service']\n",
      "['No internet service']\n",
      "['No internet service']\n"
     ]
    }
   ],
   "source": [
    "# Check for unique values in the given columns. Should only print 'No internet service' to confirm our assumption.\n",
    "print(no_www['OnlineSecurity'].unique())\n",
    "print(no_www['OnlineBackup'].unique())\n",
    "print(no_www['DeviceProtection'].unique())\n",
    "print(no_www['TechSupport'].unique())\n",
    "print(no_www['StreamingTV'].unique())\n",
    "print(no_www['StreamingMovies'].unique())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Confirmed! So, for the above columns/variables, we'll convert to binary. Which, in this case, simply means that we are removing a 3rd response option, leaving us with only 'Yes' or 'No'."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a list for looping through the dataframe and converting to binary \n",
    "binaries = ['OnlineSecurity', 'OnlineBackup', 'DeviceProtection', 'TechSupport', 'StreamingTV', 'StreamingMovies']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Loop and replace\n",
    "for i in binaries:\n",
    "    data2[i] = data2[i].replace({'No internet service' : 'No'})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, just a bit of manipulation for the sake of aesthetic in our charts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "data2[\"SeniorCitizen\"] = data2[\"SeniorCitizen\"].replace({1:\"Yes\",0:\"No\"}) # We will convert this back later - this is only for the EDA."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's look at customer churning against the different variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x1296 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_palette(sns.color_palette(col_scheme))\n",
    "fig, ax=plt.subplots(3,4,figsize=(25,18))\n",
    "\n",
    "sns.countplot(x=data2['gender'], hue=data2['Churn'], ax=ax[0,0])\n",
    "ax[0,0].set_title('Churn vs. gender')\n",
    "sns.countplot(x=data2['SeniorCitizen'], hue=data2['Churn'], ax=ax[0,1])\n",
    "ax[0,1].set_title('Churn vs. SeniorCitizen')\n",
    "sns.countplot(x=data2['Partner'], hue=data2['Churn'], ax=ax[0,2])\n",
    "ax[0,2].set_title('Churn vs. Partner')\n",
    "sns.countplot(x=data2['Dependents'], hue=data2['Churn'], ax=ax[0,3])\n",
    "ax[0,3].set_title('Churn vs. Dependents')\n",
    "\n",
    "sns.countplot(x=data2['PhoneService'], hue=data2['Churn'], ax=ax[1,0])\n",
    "ax[1,0].set_title('Churn vs. PhoneService')\n",
    "sns.countplot(x=data2['MultipleLines'], hue=data2['Churn'], ax=ax[1,1])\n",
    "ax[1,1].set_title('Churn vs. MultipleLines')\n",
    "sns.countplot(x=data2['InternetService'], hue=data2['Churn'], ax=ax[1,2])\n",
    "ax[1,2].set_title('Churn vs. InternetService')\n",
    "sns.countplot(x=data2['OnlineSecurity'], hue=data2['Churn'], ax=ax[1,3])\n",
    "ax[1,3].set_title('Churn vs. OnlineSecurity')\n",
    "\n",
    "sns.countplot(x=data2['OnlineBackup'], hue=data2['Churn'], ax=ax[2,0])\n",
    "ax[2,0].set_title('Churn vs. OnlineBackup')\n",
    "sns.countplot(x=data2['DeviceProtection'], hue=data2['Churn'], ax=ax[2,1])\n",
    "ax[2,1].set_title('Churn vs. DeviceProtection')\n",
    "sns.countplot(x=data2['TechSupport'], hue=data2['Churn'], ax=ax[2,2])\n",
    "ax[2,2].set_title('Churn vs. TechSupport')\n",
    "sns.countplot(x=data2['StreamingTV'], hue=data2['Churn'], ax=ax[2,3])\n",
    "ax[2,3].set_title('Churn vs. StreamingTV')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Only a few things of note from these graphs above:\n",
    "* Senior Citizen customers luckily make up a small portion of the business because almost as many left as stayed.\n",
    "* The amount of customers with a partner versus with no partner is fairly even. But if we look at customers who don't have a partner and left the company, the amount is almost double those with a partner who left.\n",
    "* There are less customers using Fiber Optic **InternetService** than there are using DSL service. However, the amount of customers who left the company and had Fiber Optic **InternetService** versus the amount who left and had DSL is more than double."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,3,figsize=(18,6))\n",
    "\n",
    "sns.countplot(x=data2['StreamingMovies'], hue=data2['Churn'], ax=ax[0])\n",
    "ax[0].set_title('Churn vs. StreamingMovies')\n",
    "sns.countplot(x=data2['Contract'], hue=data2['Churn'], ax=ax[1])\n",
    "ax[1].set_title('Churn vs. Contract')\n",
    "sns.countplot(x=data2['PaperlessBilling'], hue=data2['Churn'], ax=ax[2])\n",
    "ax[2].set_title('Churn vs. PaperlessBilling')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Two things stand out from these three graphs above:\n",
    "* Look at all those Month-to-month customers who leave the company in the Churn vs. Contract graph! \n",
    "* Customers with paperless billing left at a much higher frequency than customers who recieve paper copies of their bill."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,7))\n",
    "sns.countplot(x=data2['PaymentMethod'], hue=data2['Churn']).set_title('Churn vs. PaymentMethod')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is interesting: People pay their bill equally porportional across all methods, but for some reason those who pay electronically are leaving the company at almost 4 times the rate. The raw numbers are below."
   ]
  },
  {
   "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></th>\n",
       "      <th>Churn</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PaymentMethod</th>\n",
       "      <th>Churn</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Bank transfer (automatic)</th>\n",
       "      <th>No</th>\n",
       "      <td>1284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Yes</th>\n",
       "      <td>258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Credit card (automatic)</th>\n",
       "      <th>No</th>\n",
       "      <td>1289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Yes</th>\n",
       "      <td>232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Electronic check</th>\n",
       "      <th>No</th>\n",
       "      <td>1294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Yes</th>\n",
       "      <td>1071</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Mailed check</th>\n",
       "      <th>No</th>\n",
       "      <td>1296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Yes</th>\n",
       "      <td>308</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 Churn\n",
       "PaymentMethod             Churn       \n",
       "Bank transfer (automatic) No      1284\n",
       "                          Yes      258\n",
       "Credit card (automatic)   No      1289\n",
       "                          Yes      232\n",
       "Electronic check          No      1294\n",
       "                          Yes     1071\n",
       "Mailed check              No      1296\n",
       "                          Yes      308"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pay_meth = pd.DataFrame(data2.groupby(['PaymentMethod', 'Churn'])['Churn'].count())\n",
    "pay_meth"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## III. Data Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### More libraries:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.compose import ColumnTransformer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The goal in this section, Data Preprocessing, is to get our data ready for modeling. In this project we will do that in three ways:\n",
    "1. We'll be creating dummy variables using `get_dummies()` from `pandas`. This will create new columns out of our categorical features. To demonstrate, Contract will get binary columns for each response: Contract_Month-to-month : 1 or 0, Contract_One year : 1 or 0, etc. and of course, 1 or 0 will be indicate Yes or No, respectively. Another (faster) way we could've done this is using a pipline with a column transformer. However, retaining the column names to learn about feature importances would be a hefty task. \n",
    "\n",
    "2. We don't want to go creating a ton of dummy variables, so we'll only use `get_dummies()` on the categorical features that have more than 2 responses. For the rest - the binary features - we can just map all the 'Yes' responses to equal 1 and all the 'No' responses to equal 0.\n",
    "\n",
    "3. We will then take our continuous variables (**tenure**, **MonthlyCharges**, **TotalCharges**) and scale them using `StandardScaler()`. There is plenty to read on the internet about [Feature Scaling](https://en.wikipedia.org/wiki/Feature_scaling). Here is the short version: each value in a column is subtracted by the mean of the column, that number is then divided by the standard deviation of the column, that value now replaces the original value and the end result is that the new mean of the column is 0 and the new standard deviation is 1. The point of this process is to get our variables on the same scale for more effective learning. **Random Forest won't see much of an effect from this**.\n",
    "\n",
    "4. Before we do all that we will split the data into training and testing sets. The testing set will consist of 15% of the full dataset, which means that the training set will have 85% of the full dataset.\n",
    "\n",
    "5. The final thing that we want to look into is the imbalance of our data. For the sake of our sanity, we'll run the models with the data after we conduct all of the above preprocessing steps. When we get all those results, we will use *Synthetic Minority Over-Sampling Technique* (SMOTE) to balance out the data and see if we get any improvements in accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop 'customerID'\n",
    "data2.drop('customerID', axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# List of categorical and continuous features, sperating binaries, and exlcuding 'gender' and target variable 'Churn'.\n",
    "cont_feats = ['tenure', 'MonthlyCharges', 'TotalCharges']\n",
    "binary_feats = [i for i in data2.columns[:-1] if len(data2[i].unique()) == 2]\n",
    "binary_feats.remove('gender')\n",
    "cat_feats = [i for i in data2.columns[:-1] if i not in cont_feats + binary_feats]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Map the binaries 'Yes' and 'No' responses to equal 1 and 0, respectively.\n",
    "data2[binary_feats] = data2[binary_feats].replace(to_replace=['Yes', 'No'], value=[1, 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creat the dummy variables for features with more than 2 responses\n",
    "final_dat = pd.get_dummies(data2, columns=cat_feats)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train/Test Split\n",
    "train_set, test_set = train_test_split(final_dat, test_size=0.15, random_state=2021)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Finalize our list of features\n",
    "feats = [i for i in final_dat if i != 'Churn']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Finalize the structure\n",
    "data_tr = train_set[feats].copy()\n",
    "train_labels = train_set['Churn'].copy()\n",
    "data_te = test_set[feats].copy()\n",
    "test_labels = test_set['Churn'].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scale the train data\n",
    "col_trans = ColumnTransformer([\n",
    "    ('num', StandardScaler(),  cont_feats)], remainder='passthrough')\n",
    "\n",
    "train_prep = col_trans.fit_transform(data_tr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scale the test data based on the training data\n",
    "test_prep = col_trans.transform(data_te)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Restructure the list of feature names so that we can use it for feature importances later\n",
    "new_order = [i for i in data_tr.columns if i not in cont_feats]\n",
    "new_order = cont_feats+new_order"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5977, 29)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_prep.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## IV. Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Regarding metrics and scoring the performance of our model: Obviously, the company wants to retain as many customers as possible. The goal of this project is to correctly predict which customers will churn. From a business viewpoint, we would want to allocate resources toward focusing on the characteristics of a customer who churns, and spend money there with the goal of reducing churning. Interpeting that from a machine learning perspective: we want to focus on **Recall** which is a measure used for the effort of reducing False Negatives. We can't let customers slip away because we incorrectly predicted that they would stay.\n",
    "* TL;DR - We will be using `recall` for our scoring measure."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### And even more libraries:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import recall_score\n",
    "from sklearn.metrics import classification_report\n",
    "from numpy import random"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## IV.i Random Forest"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We are going to implement **RandomizedSearchCV** to help us tune our model so we are using the best parameters from a list of different parameters we would like to train on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Outline parameters to test\n",
    "param_grid = {\n",
    "    'n_estimators': random.randint(100,600,3),\n",
    "    'max_features' : random.randint(10,29,3),\n",
    "    'max_depth' : [3,4,5],\n",
    "    'bootstrap' : [True, False]\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, estimator=RandomForestClassifier(random_state=2021),\n",
       "                   param_distributions={'bootstrap': [True, False],\n",
       "                                        'max_depth': [3, 4, 5],\n",
       "                                        'max_features': array([28, 17, 23]),\n",
       "                                        'n_estimators': array([323, 392, 485])},\n",
       "                   return_train_score=True, scoring='accuracy')"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_clf = RandomForestClassifier(random_state=2021)\n",
    "rf_grid = RandomizedSearchCV(rf_clf, param_grid, cv=3, scoring='accuracy',\n",
    "                           return_train_score=True)\n",
    "rf_grid.fit(train_prep, train_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "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>Train_Error</th>\n",
       "      <th>Test_Error</th>\n",
       "      <th>Parameter_Set</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.792789</td>\n",
       "      <td>0.791031</td>\n",
       "      <td>{'n_estimators': 485, 'max_features': 17, 'max_depth': 3, 'bootstrap': True}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.804166</td>\n",
       "      <td>0.789525</td>\n",
       "      <td>{'n_estimators': 392, 'max_features': 28, 'max_depth': 5, 'bootstrap': False}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.800987</td>\n",
       "      <td>0.795381</td>\n",
       "      <td>{'n_estimators': 323, 'max_features': 17, 'max_depth': 4, 'bootstrap': True}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.815878</td>\n",
       "      <td>0.803413</td>\n",
       "      <td>{'n_estimators': 392, 'max_features': 23, 'max_depth': 5, 'bootstrap': True}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.792120</td>\n",
       "      <td>0.788856</td>\n",
       "      <td>{'n_estimators': 392, 'max_features': 28, 'max_depth': 3, 'bootstrap': False}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.792454</td>\n",
       "      <td>0.790864</td>\n",
       "      <td>{'n_estimators': 485, 'max_features': 23, 'max_depth': 3, 'bootstrap': True}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.813954</td>\n",
       "      <td>0.800904</td>\n",
       "      <td>{'n_estimators': 485, 'max_features': 28, 'max_depth': 5, 'bootstrap': True}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.797808</td>\n",
       "      <td>0.792035</td>\n",
       "      <td>{'n_estimators': 392, 'max_features': 28, 'max_depth': 4, 'bootstrap': False}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.792789</td>\n",
       "      <td>0.790362</td>\n",
       "      <td>{'n_estimators': 485, 'max_features': 28, 'max_depth': 3, 'bootstrap': True}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.801823</td>\n",
       "      <td>0.794712</td>\n",
       "      <td>{'n_estimators': 323, 'max_features': 23, 'max_depth': 4, 'bootstrap': True}</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Train_Error  Test_Error  \\\n",
       "0     0.792789    0.791031   \n",
       "1     0.804166    0.789525   \n",
       "2     0.800987    0.795381   \n",
       "3     0.815878    0.803413   \n",
       "4     0.792120    0.788856   \n",
       "5     0.792454    0.790864   \n",
       "6     0.813954    0.800904   \n",
       "7     0.797808    0.792035   \n",
       "8     0.792789    0.790362   \n",
       "9     0.801823    0.794712   \n",
       "\n",
       "                                                                   Parameter_Set  \n",
       "0   {'n_estimators': 485, 'max_features': 17, 'max_depth': 3, 'bootstrap': True}  \n",
       "1  {'n_estimators': 392, 'max_features': 28, 'max_depth': 5, 'bootstrap': False}  \n",
       "2   {'n_estimators': 323, 'max_features': 17, 'max_depth': 4, 'bootstrap': True}  \n",
       "3   {'n_estimators': 392, 'max_features': 23, 'max_depth': 5, 'bootstrap': True}  \n",
       "4  {'n_estimators': 392, 'max_features': 28, 'max_depth': 3, 'bootstrap': False}  \n",
       "5   {'n_estimators': 485, 'max_features': 23, 'max_depth': 3, 'bootstrap': True}  \n",
       "6   {'n_estimators': 485, 'max_features': 28, 'max_depth': 5, 'bootstrap': True}  \n",
       "7  {'n_estimators': 392, 'max_features': 28, 'max_depth': 4, 'bootstrap': False}  \n",
       "8   {'n_estimators': 485, 'max_features': 28, 'max_depth': 3, 'bootstrap': True}  \n",
       "9   {'n_estimators': 323, 'max_features': 23, 'max_depth': 4, 'bootstrap': True}  "
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cvres = rf_grid.cv_results_\n",
    "grid_rf_df = pd.DataFrame({\n",
    "    \"Train_Error\": cvres[\"mean_train_score\"],\n",
    "    \"Test_Error\": cvres[\"mean_test_score\"],\n",
    "    \"Parameter_Set\": cvres[\"params\"],\n",
    "}).sort_index()\n",
    "pd.set_option('display.max_colwidth', None)\n",
    "grid_rf_df.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_model = rf_grid.best_estimator_\n",
    "rf_preds = final_model.predict(test_prep)\n",
    "rfmod_pred_score = accuracy_score(test_labels, rf_preds)\n",
    "rfmod_recall = recall_score(test_labels, rf_preds, pos_label='Yes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total predictions = 1055\n"
     ]
    }
   ],
   "source": [
    "mat = confusion_matrix(test_labels, rf_preds)\n",
    "sns.heatmap(mat.T, square=True, annot=True, fmt='d', cbar=False, cmap='RdBu')\n",
    "plt.xlabel('Actual Churn', size=15)\n",
    "plt.ylabel('Predicted Churn', size=15)\n",
    "plt.show()\n",
    "print('Total predictions = %d' % len(test_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train error rate = 19.66%\n",
      "Test error rate = 20.95%\n",
      "Test Recall = 50.00%\n"
     ]
    }
   ],
   "source": [
    "print('Train error rate = %.2f%%' % ((1 - rf_grid.best_score_) * 100))\n",
    "print('Test error rate = %.2f%%' % ((1 - rfmod_pred_score) * 100))\n",
    "print('Test Recall = %.2f%%' % (rfmod_recall * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The training error rate being within ~1.3% of the testing error rate lets us know that if we have any issues modeling, it might be underfitting, but that number isn't enough to be concerned about at this time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          No       0.83      0.90      0.86       769\n",
      "         Yes       0.65      0.50      0.56       286\n",
      "\n",
      "    accuracy                           0.79      1055\n",
      "   macro avg       0.74      0.70      0.71      1055\n",
      "weighted avg       0.78      0.79      0.78      1055\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_labels, rf_preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The plot above is called a confusion matrix. The X axis is the actual value from the data set, the y axis is the predicted value. Here is what the confusion matrix tells us:\n",
    "* Our recall score, 50%, is not good. It says that we are only correctly predicting when a customer will churn half of the time. \n",
    "* Overall accuracy at 79% is also not good. We'll have to check other models and see if we can improve.\n",
    "\n",
    "The model does fairly well in predicting that customers will stay, but not so much in predicting that a customer will leave. This is a given considering the imbalance in our data favors customers who do not churn.\n",
    "\n",
    "Next, we'll look at which features the model considered most important for making its predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Create a pandas dataframe out of our train_prep array and add the columns in the correct order\n",
    "df = pd.DataFrame(train_prep, columns = new_order)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>Importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Contract_Month-to-month</th>\n",
       "      <td>0.423958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tenure</th>\n",
       "      <td>0.172991</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService_Fiber optic</th>\n",
       "      <td>0.139886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TotalCharges</th>\n",
       "      <td>0.079468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <td>0.070162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PaymentMethod_Electronic check</th>\n",
       "      <td>0.028483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService_No</th>\n",
       "      <td>0.010645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService_DSL</th>\n",
       "      <td>0.008616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TechSupport</th>\n",
       "      <td>0.007073</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SeniorCitizen</th>\n",
       "      <td>0.006717</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                Importance\n",
       "Contract_Month-to-month           0.423958\n",
       "tenure                            0.172991\n",
       "InternetService_Fiber optic       0.139886\n",
       "TotalCharges                      0.079468\n",
       "MonthlyCharges                    0.070162\n",
       "PaymentMethod_Electronic check    0.028483\n",
       "InternetService_No                0.010645\n",
       "InternetService_DSL               0.008616\n",
       "TechSupport                       0.007073\n",
       "SeniorCitizen                     0.006717"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_importances = pd.DataFrame(rf_grid.best_estimator_.feature_importances_,\n",
    "                                   index = df.columns,\n",
    "                                    columns=['Importance']).sort_values('Importance', ascending=False)\n",
    "\n",
    "rf_importances[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have a lot of features, so lets just plot the top 5."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "top_5 = rf_importances[rf_importances['Importance'] > .05]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/seaborn/_decorators.py:36: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(13,8))\n",
    "ax = sns.barplot(top_5.index, top_5['Importance'])\n",
    "ax.set_xticklabels(ax.get_xticklabels(), rotation=45, fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The barplot above tells us that the most important feature in making predictions is **Contract_Month-to-month**, followed by **tenure**, **InternetService_Fiber optic**, **TotalCharges**, and **MonthlyCharges**, respectively."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## IV.ii Logistic Regression Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now lets play with Logistic Regression and see if we get any better performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Outline parameters to test\n",
    "param_grid = [\n",
    "    {'max_iter': random.randint(100,600,3), 'C': [.1,.5,1], 'penalty': ['l2'], 'solver': ['lbfgs'], 'tol': [.01, .001, .0001]},\n",
    "    {'max_iter': random.randint(100,600,3), 'penalty': ['none'], 'tol': [.01, .001, .0001]},  \n",
    "    {'max_iter': random.randint(100,600,3), 'C': [.1,.5,1], 'penalty': ['l1'], 'solver': ['liblinear'], 'tol': [.01, .001, .0001]},\n",
    "    {'max_iter': random.randint(100,600,3), 'C': [.1,.5,1], 'penalty': ['elasticnet'], 'solver': ['saga'], 'l1_ratio':[.25, .5, .75], 'tol': [.01, .001, .0001]}\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, estimator=LogisticRegression(),\n",
       "                   param_distributions=[{'C': [0.1, 0.5, 1],\n",
       "                                         'max_iter': array([379, 294, 475]),\n",
       "                                         'penalty': ['l2'], 'solver': ['lbfgs'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]},\n",
       "                                        {'max_iter': array([386, 517, 110]),\n",
       "                                         'penalty': ['none'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]},\n",
       "                                        {'C': [0.1, 0.5, 1],\n",
       "                                         'max_iter': array([359, 327, 565]),\n",
       "                                         'penalty': ['l1'],\n",
       "                                         'solver': ['liblinear'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]},\n",
       "                                        {'C': [0.1, 0.5, 1],\n",
       "                                         'l1_ratio': [0.25, 0.5, 0.75],\n",
       "                                         'max_iter': array([361, 289, 127]),\n",
       "                                         'penalty': ['elasticnet'],\n",
       "                                         'solver': ['saga'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]}],\n",
       "                   return_train_score=True, scoring='accuracy')"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "log_clf = LogisticRegression()\n",
    "log_grid = RandomizedSearchCV(log_clf, param_grid, cv=3, scoring='accuracy',\n",
    "                           return_train_score=True)\n",
    "log_grid.fit(train_prep, train_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "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>Train_Error</th>\n",
       "      <th>Test_Error</th>\n",
       "      <th>Parameter_Set</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.805421</td>\n",
       "      <td>0.801741</td>\n",
       "      <td>{'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 361, 'l1_ratio': 0.75, 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.804417</td>\n",
       "      <td>0.803748</td>\n",
       "      <td>{'tol': 0.0001, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 289, 'l1_ratio': 0.75, 'C': 0.5}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.804919</td>\n",
       "      <td>0.803748</td>\n",
       "      <td>{'tol': 0.0001, 'solver': 'lbfgs', 'penalty': 'l2', 'max_iter': 475, 'C': 1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.804919</td>\n",
       "      <td>0.803748</td>\n",
       "      <td>{'tol': 0.0001, 'solver': 'lbfgs', 'penalty': 'l2', 'max_iter': 294, 'C': 1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.804919</td>\n",
       "      <td>0.802745</td>\n",
       "      <td>{'tol': 0.0001, 'solver': 'liblinear', 'penalty': 'l1', 'max_iter': 565, 'C': 0.5}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.805170</td>\n",
       "      <td>0.804752</td>\n",
       "      <td>{'tol': 0.001, 'penalty': 'none', 'max_iter': 386}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.804417</td>\n",
       "      <td>0.802577</td>\n",
       "      <td>{'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 289, 'l1_ratio': 0.25, 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.804919</td>\n",
       "      <td>0.803414</td>\n",
       "      <td>{'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 127, 'l1_ratio': 0.75, 'C': 1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.805086</td>\n",
       "      <td>0.801406</td>\n",
       "      <td>{'tol': 0.0001, 'solver': 'lbfgs', 'penalty': 'l2', 'max_iter': 475, 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.804919</td>\n",
       "      <td>0.803748</td>\n",
       "      <td>{'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 289, 'l1_ratio': 0.5, 'C': 1}</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Train_Error  Test_Error  \\\n",
       "0     0.805421    0.801741   \n",
       "1     0.804417    0.803748   \n",
       "2     0.804919    0.803748   \n",
       "3     0.804919    0.803748   \n",
       "4     0.804919    0.802745   \n",
       "5     0.805170    0.804752   \n",
       "6     0.804417    0.802577   \n",
       "7     0.804919    0.803414   \n",
       "8     0.805086    0.801406   \n",
       "9     0.804919    0.803748   \n",
       "\n",
       "                                                                                             Parameter_Set  \n",
       "0    {'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 361, 'l1_ratio': 0.75, 'C': 0.1}  \n",
       "1  {'tol': 0.0001, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 289, 'l1_ratio': 0.75, 'C': 0.5}  \n",
       "2                             {'tol': 0.0001, 'solver': 'lbfgs', 'penalty': 'l2', 'max_iter': 475, 'C': 1}  \n",
       "3                             {'tol': 0.0001, 'solver': 'lbfgs', 'penalty': 'l2', 'max_iter': 294, 'C': 1}  \n",
       "4                       {'tol': 0.0001, 'solver': 'liblinear', 'penalty': 'l1', 'max_iter': 565, 'C': 0.5}  \n",
       "5                                                       {'tol': 0.001, 'penalty': 'none', 'max_iter': 386}  \n",
       "6    {'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 289, 'l1_ratio': 0.25, 'C': 0.1}  \n",
       "7      {'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 127, 'l1_ratio': 0.75, 'C': 1}  \n",
       "8                           {'tol': 0.0001, 'solver': 'lbfgs', 'penalty': 'l2', 'max_iter': 475, 'C': 0.1}  \n",
       "9       {'tol': 0.01, 'solver': 'saga', 'penalty': 'elasticnet', 'max_iter': 289, 'l1_ratio': 0.5, 'C': 1}  "
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cvres = log_grid.cv_results_\n",
    "grid_log_df = pd.DataFrame({\n",
    "    \"Train_Error\": cvres[\"mean_train_score\"],\n",
    "    \"Test_Error\": cvres[\"mean_test_score\"],\n",
    "    \"Parameter_Set\": cvres[\"params\"],\n",
    "}).sort_index()\n",
    "grid_log_df.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_model = log_grid.best_estimator_\n",
    "log_preds = final_model.predict(test_prep)\n",
    "logmod_pred_score = accuracy_score(test_labels, log_preds)\n",
    "logmod_recall = recall_score(test_labels, log_preds, pos_label='Yes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total predictions = 1055\n"
     ]
    }
   ],
   "source": [
    "mat = confusion_matrix(test_labels, log_preds)\n",
    "sns.heatmap(mat.T, square=True, annot=True, fmt='d', cbar=False, cmap='RdBu')\n",
    "plt.xlabel('Actual Churn', size=15)\n",
    "plt.ylabel('Predicted Churn', size=15)\n",
    "plt.show()\n",
    "print('Total predictions = %d' % len(test_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train error rate = 19.52%\n",
      "Test error rate = 19.62%\n",
      "Test Recall = 54.90%\n"
     ]
    }
   ],
   "source": [
    "print('Train error rate = %.2f%%' % ((1 - log_grid.best_score_) * 100))\n",
    "print('Test error rate = %.2f%%' % ((1 - logmod_pred_score) * 100))\n",
    "print('Test Recall = %.2f%%' % (logmod_recall * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This model is even tighter than Random Forest, we have almost identical testing and training error rates."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          No       0.84      0.90      0.87       769\n",
      "         Yes       0.67      0.55      0.60       286\n",
      "\n",
      "    accuracy                           0.80      1055\n",
      "   macro avg       0.76      0.72      0.74      1055\n",
      "weighted avg       0.80      0.80      0.80      1055\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_labels, log_preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Understanding the Logistic Regression results:\n",
    "* Very subtle, ever so slight, but still, improvement on our focus metric - recall: 54.90%\n",
    "* We're still in the same ballpark for overall performance\n",
    "* To reiterate, we improved slightly on the recall score, but considering the way this project seems to be playing out, we are going to have to be happy with any imporovement"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking at feature importanes for this model is a bit of a different venture. We will look at the coefficient values for our features, square them so that they are all positive and then take the square root, and rank them in descending order. The coefficients basically tell us how much weight each feature carries in predicting whether or not a customer will churn."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(log_grid.best_estimator_.coef_, columns = new_order).transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>square</th>\n",
       "      <th>sqrt</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>InternetService_No</th>\n",
       "      <td>-2.551594</td>\n",
       "      <td>6.510632</td>\n",
       "      <td>2.551594</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService_Fiber optic</th>\n",
       "      <td>1.730275</td>\n",
       "      <td>2.993853</td>\n",
       "      <td>1.730275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <td>-1.640833</td>\n",
       "      <td>2.692334</td>\n",
       "      <td>1.640833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tenure</th>\n",
       "      <td>-1.392298</td>\n",
       "      <td>1.938493</td>\n",
       "      <td>1.392298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Contract_Two year</th>\n",
       "      <td>-1.106902</td>\n",
       "      <td>1.225232</td>\n",
       "      <td>1.106902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MultipleLines_No phone service</th>\n",
       "      <td>-1.016609</td>\n",
       "      <td>1.033493</td>\n",
       "      <td>1.016609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>StreamingTV</th>\n",
       "      <td>0.807969</td>\n",
       "      <td>0.652814</td>\n",
       "      <td>0.807969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>StreamingMovies</th>\n",
       "      <td>0.709934</td>\n",
       "      <td>0.504007</td>\n",
       "      <td>0.709934</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TotalCharges</th>\n",
       "      <td>0.651526</td>\n",
       "      <td>0.424486</td>\n",
       "      <td>0.651526</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gender_Male</th>\n",
       "      <td>-0.625453</td>\n",
       "      <td>0.391192</td>\n",
       "      <td>0.625453</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                       0    square      sqrt\n",
       "InternetService_No             -2.551594  6.510632  2.551594\n",
       "InternetService_Fiber optic     1.730275  2.993853  1.730275\n",
       "MonthlyCharges                 -1.640833  2.692334  1.640833\n",
       "tenure                         -1.392298  1.938493  1.392298\n",
       "Contract_Two year              -1.106902  1.225232  1.106902\n",
       "MultipleLines_No phone service -1.016609  1.033493  1.016609\n",
       "StreamingTV                     0.807969  0.652814  0.807969\n",
       "StreamingMovies                 0.709934  0.504007  0.709934\n",
       "TotalCharges                    0.651526  0.424486  0.651526\n",
       "gender_Male                    -0.625453  0.391192  0.625453"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['square'] = df[0] ** 2\n",
    "df['sqrt'] = [np.sqrt(i) for i in df['square']]\n",
    "df = df.sort_values('sqrt', ascending=False)\n",
    "df[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "top5 = df[df['sqrt'] > 1.1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/seaborn/_decorators.py:36: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(13,8))\n",
    "ax = sns.barplot(top5.index, top5['sqrt'])\n",
    "ax.set_xticklabels(ax.get_xticklabels(), rotation=45, fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Logistic Regression has a slightly different take. It says that **InternetService_No** is the most important feature, followed by **InternetSevice_Fiber optic**, **MonthlyCharges**, **tenure**, and **Contract_Two year**."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## IV.iii Support Vector Machine Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "param_grid = [    \n",
    "    {'C': [.1,1,10], 'max_iter': random.randint(1000,3000,3), 'tol': [1e-4, 1e-10], 'gamma' : [3,5,7]},\n",
    "    {'C': [.1,1,10], 'max_iter': [-1], 'tol': [1e-4, 1e-10]},\n",
    "    {'C': [.1,1,10], 'max_iter': random.randint(1000,3000,3), 'kernel': ['linear']}\n",
    "  ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2704).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2704).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2704).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1613).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1613).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1613).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1473).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1473).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1473).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2704).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2704).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2704).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2488).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1684).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1473).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, estimator=SVC(random_state=2021), n_iter=15,\n",
       "                   param_distributions=[{'C': [0.1, 1, 10], 'gamma': [3, 5, 7],\n",
       "                                         'max_iter': array([2704, 1684, 2488]),\n",
       "                                         'tol': [0.0001, 1e-10]},\n",
       "                                        {'C': [0.1, 1, 10], 'max_iter': [-1],\n",
       "                                         'tol': [0.0001, 1e-10]},\n",
       "                                        {'C': [0.1, 1, 10],\n",
       "                                         'kernel': ['linear'],\n",
       "                                         'max_iter': array([1473, 1613, 2801])}],\n",
       "                   return_train_score=True, scoring='accuracy')"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Numerous convergence warnings\n",
    "svm_clf = SVC(random_state=2021)\n",
    "svm_grid = RandomizedSearchCV(svm_clf, param_grid, cv=3,\n",
    "                           scoring='accuracy',\n",
    "                           return_train_score=True, n_iter=15)\n",
    "svm_grid.fit(train_prep, train_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "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>Train_Error</th>\n",
       "      <th>Test_Error</th>\n",
       "      <th>Parameter_Set</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.735151</td>\n",
       "      <td>0.735151</td>\n",
       "      <td>{'tol': 1e-10, 'max_iter': 2488, 'gamma': 5, 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.735151</td>\n",
       "      <td>0.735151</td>\n",
       "      <td>{'tol': 0.0001, 'max_iter': 2488, 'gamma': 3, 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.735151</td>\n",
       "      <td>0.735151</td>\n",
       "      <td>{'tol': 1e-10, 'max_iter': 2704, 'gamma': 7, 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.757236</td>\n",
       "      <td>0.752720</td>\n",
       "      <td>{'max_iter': 1613, 'kernel': 'linear', 'C': 1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.801991</td>\n",
       "      <td>0.798728</td>\n",
       "      <td>{'max_iter': 1473, 'kernel': 'linear', 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.961686</td>\n",
       "      <td>0.754392</td>\n",
       "      <td>{'tol': 0.0001, 'max_iter': 1684, 'gamma': 5, 'C': 1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.961184</td>\n",
       "      <td>0.755228</td>\n",
       "      <td>{'tol': 1e-10, 'max_iter': 2488, 'gamma': 5, 'C': 1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.974235</td>\n",
       "      <td>0.756399</td>\n",
       "      <td>{'tol': 1e-10, 'max_iter': 1684, 'gamma': 3, 'C': 10}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.973063</td>\n",
       "      <td>0.756399</td>\n",
       "      <td>{'tol': 0.0001, 'max_iter': 2704, 'gamma': 3, 'C': 10}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.975322</td>\n",
       "      <td>0.752551</td>\n",
       "      <td>{'tol': 0.0001, 'max_iter': 2488, 'gamma': 5, 'C': 10}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.735151</td>\n",
       "      <td>0.735151</td>\n",
       "      <td>{'tol': 1e-10, 'max_iter': 1684, 'gamma': 7, 'C': 0.1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.872009</td>\n",
       "      <td>0.782833</td>\n",
       "      <td>{'tol': 0.0001, 'max_iter': -1, 'C': 10}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.964949</td>\n",
       "      <td>0.754225</td>\n",
       "      <td>{'tol': 1e-10, 'max_iter': 1684, 'gamma': 7, 'C': 1}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.975991</td>\n",
       "      <td>0.752049</td>\n",
       "      <td>{'tol': 0.0001, 'max_iter': 2488, 'gamma': 7, 'C': 10}</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.974235</td>\n",
       "      <td>0.756399</td>\n",
       "      <td>{'tol': 0.0001, 'max_iter': 1684, 'gamma': 3, 'C': 10}</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Train_Error  Test_Error  \\\n",
       "0      0.735151    0.735151   \n",
       "1      0.735151    0.735151   \n",
       "2      0.735151    0.735151   \n",
       "3      0.757236    0.752720   \n",
       "4      0.801991    0.798728   \n",
       "5      0.961686    0.754392   \n",
       "6      0.961184    0.755228   \n",
       "7      0.974235    0.756399   \n",
       "8      0.973063    0.756399   \n",
       "9      0.975322    0.752551   \n",
       "10     0.735151    0.735151   \n",
       "11     0.872009    0.782833   \n",
       "12     0.964949    0.754225   \n",
       "13     0.975991    0.752049   \n",
       "14     0.974235    0.756399   \n",
       "\n",
       "                                              Parameter_Set  \n",
       "0    {'tol': 1e-10, 'max_iter': 2488, 'gamma': 5, 'C': 0.1}  \n",
       "1   {'tol': 0.0001, 'max_iter': 2488, 'gamma': 3, 'C': 0.1}  \n",
       "2    {'tol': 1e-10, 'max_iter': 2704, 'gamma': 7, 'C': 0.1}  \n",
       "3            {'max_iter': 1613, 'kernel': 'linear', 'C': 1}  \n",
       "4          {'max_iter': 1473, 'kernel': 'linear', 'C': 0.1}  \n",
       "5     {'tol': 0.0001, 'max_iter': 1684, 'gamma': 5, 'C': 1}  \n",
       "6      {'tol': 1e-10, 'max_iter': 2488, 'gamma': 5, 'C': 1}  \n",
       "7     {'tol': 1e-10, 'max_iter': 1684, 'gamma': 3, 'C': 10}  \n",
       "8    {'tol': 0.0001, 'max_iter': 2704, 'gamma': 3, 'C': 10}  \n",
       "9    {'tol': 0.0001, 'max_iter': 2488, 'gamma': 5, 'C': 10}  \n",
       "10   {'tol': 1e-10, 'max_iter': 1684, 'gamma': 7, 'C': 0.1}  \n",
       "11                 {'tol': 0.0001, 'max_iter': -1, 'C': 10}  \n",
       "12     {'tol': 1e-10, 'max_iter': 1684, 'gamma': 7, 'C': 1}  \n",
       "13   {'tol': 0.0001, 'max_iter': 2488, 'gamma': 7, 'C': 10}  \n",
       "14   {'tol': 0.0001, 'max_iter': 1684, 'gamma': 3, 'C': 10}  "
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cvres = svm_grid.cv_results_\n",
    "grid_svc_df = pd.DataFrame({\n",
    "    \"Train_Error\": cvres[\"mean_train_score\"],\n",
    "    \"Test_Error\": cvres[\"mean_test_score\"],\n",
    "    \"Parameter_Set\": cvres[\"params\"],\n",
    "}).sort_index()\n",
    "grid_svc_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_model = svm_grid.best_estimator_\n",
    "svm_preds = final_model.predict(test_prep)\n",
    "svmmod_pred_score = accuracy_score(test_labels, svm_preds)\n",
    "svmmod_recall = recall_score(test_labels, svm_preds, pos_label='Yes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total predictions = 1055\n"
     ]
    }
   ],
   "source": [
    "mat = confusion_matrix(test_labels, svm_preds)\n",
    "sns.heatmap(mat.T, square=True, annot=True, fmt='d', cbar=False, cmap='RdBu')\n",
    "plt.xlabel('Actual Churn', size=15)\n",
    "plt.ylabel('Predicted Churn', size=15)\n",
    "plt.show()\n",
    "print('Total predictions = %d' % len(test_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train error rate = 20.13%\n",
      "Test error rate = 21.52%\n",
      "Test Recall = 50.70%\n"
     ]
    }
   ],
   "source": [
    "print('Train error rate = %.2f%%' % ((1 - svm_grid.best_score_) * 100))\n",
    "print('Test error rate = %.2f%%' % ((1 - svmmod_pred_score) * 100))\n",
    "print('Test Recall = %.2f%%' % (svmmod_recall * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Minor concern for underfitting here, but we're not going to worry about it at this point."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          No       0.83      0.89      0.86       769\n",
      "         Yes       0.63      0.51      0.56       286\n",
      "\n",
      "    accuracy                           0.78      1055\n",
      "   macro avg       0.73      0.70      0.71      1055\n",
      "weighted avg       0.77      0.78      0.78      1055\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_labels, svm_preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In reviewing the **SVM** the overall performance of the model was worst here. The metric we are concerned, recall score, didn't improve either at 50.70%."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We look at feature importane here the same way as we did with Logistic Regression."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(svm_grid.best_estimator_.coef_, columns = new_order).transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>square</th>\n",
       "      <th>sqrt</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TotalCharges</th>\n",
       "      <td>-0.765809</td>\n",
       "      <td>0.586464</td>\n",
       "      <td>0.765809</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService_Fiber optic</th>\n",
       "      <td>0.590489</td>\n",
       "      <td>0.348677</td>\n",
       "      <td>0.590489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService_DSL</th>\n",
       "      <td>-0.367819</td>\n",
       "      <td>0.135290</td>\n",
       "      <td>0.367819</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <td>0.333857</td>\n",
       "      <td>0.111460</td>\n",
       "      <td>0.333857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TechSupport</th>\n",
       "      <td>-0.332520</td>\n",
       "      <td>0.110569</td>\n",
       "      <td>0.332520</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService_No</th>\n",
       "      <td>-0.222670</td>\n",
       "      <td>0.049582</td>\n",
       "      <td>0.222670</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OnlineBackup</th>\n",
       "      <td>-0.203544</td>\n",
       "      <td>0.041430</td>\n",
       "      <td>0.203544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PaymentMethod_Electronic check</th>\n",
       "      <td>0.193118</td>\n",
       "      <td>0.037294</td>\n",
       "      <td>0.193118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PaperlessBilling</th>\n",
       "      <td>0.158202</td>\n",
       "      <td>0.025028</td>\n",
       "      <td>0.158202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PaymentMethod_Credit card (automatic)</th>\n",
       "      <td>-0.149599</td>\n",
       "      <td>0.022380</td>\n",
       "      <td>0.149599</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              0    square      sqrt\n",
       "TotalCharges                          -0.765809  0.586464  0.765809\n",
       "InternetService_Fiber optic            0.590489  0.348677  0.590489\n",
       "InternetService_DSL                   -0.367819  0.135290  0.367819\n",
       "MonthlyCharges                         0.333857  0.111460  0.333857\n",
       "TechSupport                           -0.332520  0.110569  0.332520\n",
       "InternetService_No                    -0.222670  0.049582  0.222670\n",
       "OnlineBackup                          -0.203544  0.041430  0.203544\n",
       "PaymentMethod_Electronic check         0.193118  0.037294  0.193118\n",
       "PaperlessBilling                       0.158202  0.025028  0.158202\n",
       "PaymentMethod_Credit card (automatic) -0.149599  0.022380  0.149599"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['square'] = df[0] ** 2\n",
    "df['sqrt'] = [np.sqrt(i) for i in df['square']]\n",
    "df = df.sort_values('sqrt', ascending=False)\n",
    "df[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "top5 = df[df['sqrt'] > 0.25]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/seaborn/_decorators.py:36: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(13,8))\n",
    "ax = sns.barplot(top5.index, top5['sqrt'])\n",
    "ax.set_xticklabels(ax.get_xticklabels(), rotation=45, fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**SVM** with another take on feature importances, considers **TotalCharges** most important. Then, **InternetService_Fiber optic**, **InternetService_DSL**, **MonthlyCharges**, and **TechSupport**."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### IV.iv SMOTE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's implement SMOTE and see if we get any improvements."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "smt = SMOTE(random_state = 2021) \n",
    "x_smt, y_smt = smt.fit_sample(train_prep, train_labels) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "param_grid = {\n",
    "    'n_estimators': random.randint(100,600,3),\n",
    "    'max_features' : random.randint(10,29,3),\n",
    "    'max_depth' : [3,4,5],\n",
    "    'bootstrap' : [True, False]\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, estimator=RandomForestClassifier(random_state=2021),\n",
       "                   param_distributions={'bootstrap': [True, False],\n",
       "                                        'max_depth': [3, 4, 5],\n",
       "                                        'max_features': array([25, 22, 17]),\n",
       "                                        'n_estimators': array([530, 129, 459])},\n",
       "                   scoring='accuracy')"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_clf = RandomForestClassifier(random_state=2021)\n",
    "rf_grid = RandomizedSearchCV(rf_clf, param_grid, cv=3, scoring='accuracy')\n",
    "rf_grid.fit(x_smt, y_smt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_model = rf_grid.best_estimator_\n",
    "rf_preds = final_model.predict(test_prep)\n",
    "rfmod_pred_score = accuracy_score(test_labels, rf_preds)\n",
    "rfmod_recall = recall_score(test_labels, rf_preds, pos_label='Yes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train error rate = 20.50%\n",
      "Test error rate = 23.03%\n",
      "Test Recall = 76.92%\n"
     ]
    }
   ],
   "source": [
    "print('Train error rate = %.2f%%' % ((1 - rf_grid.best_score_) * 100))\n",
    "print('Test error rate = %.2f%%' % ((1 - rfmod_pred_score) * 100))\n",
    "print('Test Recall = %.2f%%' % (rfmod_recall * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          No       0.90      0.77      0.83       769\n",
      "         Yes       0.55      0.77      0.64       286\n",
      "\n",
      "    accuracy                           0.77      1055\n",
      "   macro avg       0.73      0.77      0.74      1055\n",
      "weighted avg       0.81      0.77      0.78      1055\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_labels, rf_preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Outline parameters to test\n",
    "param_grid = [\n",
    "    {'max_iter': random.randint(100,600,3), 'C': [.1,.5,1], 'penalty': ['l2'], 'solver': ['lbfgs'], 'tol': [.01, .001, .0001]},\n",
    "    {'max_iter': random.randint(100,600,3), 'penalty': ['none'], 'tol': [.01, .001, .0001]},  \n",
    "    {'max_iter': random.randint(100,600,3), 'C': [.1,.5,1], 'penalty': ['l1'], 'solver': ['liblinear'], 'tol': [.01, .001, .0001]},\n",
    "    {'max_iter': random.randint(100,600,3), 'C': [.1,.5,1], 'penalty': ['elasticnet'], 'solver': ['saga'], 'l1_ratio':[.25, .5, .75], 'tol': [.01, .001, .0001]}\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, estimator=LogisticRegression(),\n",
       "                   param_distributions=[{'C': [0.1, 0.5, 1],\n",
       "                                         'max_iter': array([331, 288, 582]),\n",
       "                                         'penalty': ['l2'], 'solver': ['lbfgs'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]},\n",
       "                                        {'max_iter': array([102, 103, 591]),\n",
       "                                         'penalty': ['none'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]},\n",
       "                                        {'C': [0.1, 0.5, 1],\n",
       "                                         'max_iter': array([398, 437, 140]),\n",
       "                                         'penalty': ['l1'],\n",
       "                                         'solver': ['liblinear'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]},\n",
       "                                        {'C': [0.1, 0.5, 1],\n",
       "                                         'l1_ratio': [0.25, 0.5, 0.75],\n",
       "                                         'max_iter': array([454, 386, 468]),\n",
       "                                         'penalty': ['elasticnet'],\n",
       "                                         'solver': ['saga'],\n",
       "                                         'tol': [0.01, 0.001, 0.0001]}],\n",
       "                   scoring='accuracy')"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "log_clf = LogisticRegression()\n",
    "log_grid = RandomizedSearchCV(log_clf, param_grid, cv=3, scoring='accuracy')\n",
    "log_grid.fit(x_smt, y_smt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_model = log_grid.best_estimator_\n",
    "log_preds = final_model.predict(test_prep)\n",
    "logmod_pred_score = accuracy_score(test_labels, log_preds)\n",
    "logmod_recall = recall_score(test_labels, log_preds, pos_label='Yes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train error rate = 21.96%\n",
      "Test error rate = 26.07%\n",
      "Test Recall = 75.52%\n"
     ]
    }
   ],
   "source": [
    "print('Train error rate = %.2f%%' % ((1 - log_grid.best_score_) * 100))\n",
    "print('Test error rate = %.2f%%' % ((1 - logmod_pred_score) * 100))\n",
    "print('Test Recall = %.2f%%' % (logmod_recall * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          No       0.89      0.73      0.80       769\n",
      "         Yes       0.51      0.76      0.61       286\n",
      "\n",
      "    accuracy                           0.74      1055\n",
      "   macro avg       0.70      0.74      0.71      1055\n",
      "weighted avg       0.79      0.74      0.75      1055\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_labels, log_preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Support Vector Machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "param_grid = [    \n",
    "    {'C': [.1,1,10], 'max_iter': random.randint(1000,3000,3), 'tol': [1e-4, 1e-10], 'gamma' : [3,5,7]},\n",
    "    {'C': [.1,1,10], 'max_iter': [-1], 'tol': [1e-4, 1e-10]},\n",
    "    {'C': [.1,1,10], 'max_iter': random.randint(1000,3000,3), 'kernel': ['linear']}\n",
    "  ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1988).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1988).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1988).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2296).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2296).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2296).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2118).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2118).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=2118).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1032).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n",
      "/Users/lawrencelove/Library/Python/3.8/lib/python/site-packages/sklearn/svm/_base.py:246: ConvergenceWarning: Solver terminated early (max_iter=1924).  Consider pre-processing your data with StandardScaler or MinMaxScaler.\n",
      "  warnings.warn('Solver terminated early (max_iter=%i).'\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, estimator=SVC(random_state=2021), n_iter=15,\n",
       "                   param_distributions=[{'C': [0.1, 1, 10], 'gamma': [3, 5, 7],\n",
       "                                         'max_iter': array([1924, 2296, 1032]),\n",
       "                                         'tol': [0.0001, 1e-10]},\n",
       "                                        {'C': [0.1, 1, 10], 'max_iter': [-1],\n",
       "                                         'tol': [0.0001, 1e-10]},\n",
       "                                        {'C': [0.1, 1, 10],\n",
       "                                         'kernel': ['linear'],\n",
       "                                         'max_iter': array([1988, 1938, 2118])}],\n",
       "                   scoring='accuracy')"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Numerous convergence warnings\n",
    "svm_clf = SVC(random_state=2021)\n",
    "svm_grid = RandomizedSearchCV(svm_clf, param_grid, cv=3,\n",
    "                           scoring='accuracy',\n",
    "                           n_iter=15)\n",
    "svm_grid.fit(x_smt, y_smt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_model = svm_grid.best_estimator_\n",
    "svm_preds = final_model.predict(test_prep)\n",
    "svmmod_pred_score = accuracy_score(test_labels, svm_preds)\n",
    "svmmod_recall = recall_score(test_labels, svm_preds, pos_label='Yes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train error rate = 11.04%\n",
      "Test error rate = 25.02%\n",
      "Test Recall = 37.76%\n"
     ]
    }
   ],
   "source": [
    "print('Train error rate = %.2f%%' % ((1 - svm_grid.best_score_) * 100))\n",
    "print('Test error rate = %.2f%%' % ((1 - svmmod_pred_score) * 100))\n",
    "print('Test Recall = %.2f%%' % (svmmod_recall * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          No       0.79      0.89      0.84       769\n",
      "         Yes       0.56      0.38      0.45       286\n",
      "\n",
      "    accuracy                           0.75      1055\n",
      "   macro avg       0.67      0.63      0.64      1055\n",
      "weighted avg       0.73      0.75      0.73      1055\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(test_labels, svm_preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## V. Conclusions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's wrap it all up/summarize and make some conclusions. We can use all this information to raise some more questions as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_recall = [rfmod_recall, logmod_recall, svmmod_recall]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_acc = [rfmod_pred_score, logmod_pred_score, svmmod_pred_score]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(list(zip(all_recall, all_acc)), columns =['Recall', 'Accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Model'] = ['Random Forest', 'Logistic Regresssion', 'Support Vector Machine']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[['Model', 'Recall', 'Accuracy']].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Model</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Random Forest</td>\n",
       "      <td>0.769231</td>\n",
       "      <td>0.769668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Logistic Regresssion</td>\n",
       "      <td>0.755245</td>\n",
       "      <td>0.739336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Support Vector Machine</td>\n",
       "      <td>0.377622</td>\n",
       "      <td>0.749763</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    Model    Recall  Accuracy\n",
       "0           Random Forest  0.769231  0.769668\n",
       "1    Logistic Regresssion  0.755245  0.739336\n",
       "2  Support Vector Machine  0.377622  0.749763"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plt.style.use('seaborn')\n",
    "ax = df.plot(x='Model', y=['Accuracy', 'Recall'], kind=\"bar\", \n",
    "             width=.4, ylim=(0,.99), figsize=(9,7), title='Scores')\n",
    "ax.set_xticklabels(ax.get_xticklabels(), rotation=45, fontsize=12)\n",
    "plt.yticks(np.arange(0, 1, step=0.1))\n",
    "ax.set_xlabel(\"\")\n",
    "plt.legend(prop={'size': 12}, loc=1)\n",
    "plt.show()\n",
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we aggregate the feature importances, we find that our most important features are: **InternetService_Fiber optic**, **MonthlyCharges**, **TotalCharges**, and **tenure**. These features were most commonly in the top 5 of our models' feature importances.\n",
    "\n",
    "When we consider the Exploratory Data Analysis we did, it is clear that this company has some issue with their **Month-to-month** customers. What kind of incentives can this company offer to customers to get them to sign **One-year** or **Two-year** contracts? What adjustments can be made to **Month-to-Month** contracts that would be more favorable to customers, without taking away the appeal of a **One-year** or **Two-year** contract?\n",
    "\n",
    "**InternetService** was another important factor in making predictions, specifically, **Fiber optic internet service**. We are not privy to the historical timing of this data, but it is a major problem that more people with **fiber optic internet service** are leaving than with **DSL internet service**. With fiber optics being the newer technology, maybe it’s possible that the company was still working out some kinks? Regardless, if the newer (and supposed to be more efficient) technology is causing people to run from your company, you have a problem. More in-depth research is needed to fully understand this **InternetService** issue, but at the least we now know that it is an area to focus on.\n",
    "\n",
    "As far as model performance, Random Forest with SMOTE performed best. Second place goes to the Logistic Regression, and Support Vector Machine last.\n",
    "\n",
    "We were able to learn a lot from the project: how long customers are staying with the company, which features are most likely to contribute to customers leaving, and we have 3 different models to test for predicting customer churning. Most importantly, we've raised concerns that allow for some questions to be answered, for further investigation to be completed, and for more information to be gathered. We now have targets to focus in on and conduct a new study in the future."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}