Waze project lab.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "DtNBZFHO3M7n"
   },
   "source": [
    "# **Waze Project**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "kaOj1equPMAb"
   },
   "source": [
    "Your team is more than halfway through their user churn project. Earlier, you completed a project proposal, used Python to explore and analyze Waze’s user data, created data visualizations, and conducted a hypothesis test. Now, leadership wants your team to build a regression model to predict user churn based on a variety of variables.\n",
    "\n",
    "You check your inbox and discover a new email from Ursula Sayo, Waze's Operations Manager. Ursula asks your team about the details of the regression model. You also notice two follow-up emails from your supervisor, May Santner. The first email is a response to Ursula, and says that the team will build a binomial logistic regression model. In her second email, May asks you to help build the model and prepare an executive summary to share your results."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rgSbVJvomcVa"
   },
   "source": [
    "# **Course 5 End-of-course project: Regression modeling**\n",
    "\n",
    "In this activity, you will build a binomial logistic regression model. As you have learned, logistic regression helps you estimate the probability of an outcome. For data science professionals, this is a useful skill because it allows you to consider more than one variable against the variable you're measuring against. This opens the door for much more thorough and flexible analysis to be completed.\n",
    "<br/>\n",
    "\n",
    "**The purpose** of this project is to demostrate knowledge of exploratory data analysis (EDA) and a binomial logistic regression model.\n",
    "\n",
    "**The goal** is to build a binomial logistic regression model and evaluate the model's performance.\n",
    "<br/>\n",
    "\n",
    "*This activity has three parts:*\n",
    "\n",
    "**Part 1:** EDA & Checking Model Assumptions\n",
    "* What are some purposes of EDA before constructing a binomial logistic regression model?\n",
    "\n",
    "**Part 2:** Model Building and Evaluation\n",
    "* What resources do you find yourself using as you complete this stage?\n",
    "\n",
    "**Part 3:** Interpreting Model Results\n",
    "\n",
    "* What key insights emerged from your model(s)?\n",
    "\n",
    "* What business recommendations do you propose based on the models built?\n",
    "\n",
    "<br/>\n",
    "\n",
    "Follow the instructions and answer the question below to complete the activity. Then, you will complete an executive summary using the questions listed on the PACE Strategy Document.\n",
    "\n",
    "Be sure to complete this activity before moving on. The next course item will provide you with a completed exemplar to compare to your own work."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3CYH59R_fgIa"
   },
   "source": [
    "# **Build a regression model**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3UCHQclzQDUL"
   },
   "source": [
    "<img src='images/Pace.png' width='100' height='100' align=left>\n",
    "\n",
    "# **PACE stages**\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VARIxaLDjZCb"
   },
   "source": [
    "Throughout these project notebooks, you'll see references to the problem-solving framework PACE. The following notebook components are labeled with the respective PACE stage: Plan, Analyze, Construct, and Execute."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "F5O5cx_qQJmX"
   },
   "source": [
    "<img src='images/Plan.png' width='100' height='100' align=left>\n",
    "\n",
    "\n",
    "## **PACE: Plan**\n",
    "Consider the questions in your PACE Strategy Document to reflect on the Plan stage."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "D8qYlvkLQsf2"
   },
   "source": [
    "### **Task 1. Imports and data loading**\n",
    "Import the data and packages that you've learned are needed for building logistic regression models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "ccfeg6X6eOVZ"
   },
   "outputs": [],
   "source": [
    "\n",
    "# Packages for numerics + dataframes\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "# Packages for visualization\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "# Packages for Logistic Regression & Confusion Matrix\n",
    "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report, accuracy_score, precision_score, \\\n",
    "recall_score, f1_score, confusion_matrix, ConfusionMatrixDisplay\n",
    "from sklearn.linear_model import LogisticRegression\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LjljvyG32kqe"
   },
   "source": [
    "Import the dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "TyR3sBUYJBO8"
   },
   "outputs": [],
   "source": [
    "# Load the dataset by running this cell\n",
    "\n",
    "df = pd.read_csv('waze_dataset.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OnrvCSfHUWPv"
   },
   "source": [
    "## **PACE: Analyze**\n",
    "\n",
    "Consider the questions in your PACE Strategy Document to reflect on the Analyze stage.\n",
    "\n",
    "In this stage, consider the following question:\n",
    "\n",
    "\n",
    "* What are some purposes of EDA before constructing a binomial logistic regression model?\n",
    "\n",
    "> *Outliers and extreme data values can significantly impact logistic\n",
    "regression models. After visualizing data, make a plan for addressing outliers by dropping rows, substituting extreme data with average data, and/or removing data values greater than 3 standard deviations.*\n",
    "\n",
    ">  *EDA activities also include identifying missing data to help the analyst make decisions on their exclusion or inclusion by substituting values with dataset means, medians, and other similar methods.*\n",
    "\n",
    "> *Additionally, it can be useful to create variables by multiplying variables together or calculating the ratio between two variables. For example, in this dataset you can create a drives_sessions_ratio variable by dividing drives by sessions.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rIcDG2e66wt9"
   },
   "source": [
    "### **Task 2a. Explore data with EDA**\n",
    "\n",
    "Analyze and discover data, looking for correlations, missing data, potential outliers, and/or duplicates.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "offDH5p62x73"
   },
   "source": [
    "Start with `shape` and `info()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 18,
     "status": "ok",
     "timestamp": 1685661316022,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "T4Ag-sZhWg6K",
    "outputId": "d4400d88-7ff9-4266-a92a-014fbdb62251"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(14999, 13)\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 14999 entries, 0 to 14998\n",
      "Data columns (total 13 columns):\n",
      " #   Column                   Non-Null Count  Dtype  \n",
      "---  ------                   --------------  -----  \n",
      " 0   ID                       14999 non-null  int64  \n",
      " 1   label                    14299 non-null  object \n",
      " 2   sessions                 14999 non-null  int64  \n",
      " 3   drives                   14999 non-null  int64  \n",
      " 4   total_sessions           14999 non-null  float64\n",
      " 5   n_days_after_onboarding  14999 non-null  int64  \n",
      " 6   total_navigations_fav1   14999 non-null  int64  \n",
      " 7   total_navigations_fav2   14999 non-null  int64  \n",
      " 8   driven_km_drives         14999 non-null  float64\n",
      " 9   duration_minutes_drives  14999 non-null  float64\n",
      " 10  activity_days            14999 non-null  int64  \n",
      " 11  driving_days             14999 non-null  int64  \n",
      " 12  device                   14999 non-null  object \n",
      "dtypes: float64(3), int64(8), object(2)\n",
      "memory usage: 1.5+ MB\n"
     ]
    }
   ],
   "source": [
    "print(df.shape)\n",
    "\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vpTUYqFM47xy"
   },
   "source": [
    "**Question:** Are there any missing values in your data?\n",
    "\n",
    "> *Yes, the `label` column is missing 700 values.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ZCNi5-Ur2_en"
   },
   "source": [
    "Use `head()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 270
    },
    "executionInfo": {
     "elapsed": 16,
     "status": "ok",
     "timestamp": 1685661316022,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "ll2pxoClXgmx",
    "outputId": "87aaf063-308a-4081-f20f-a5eeaf98e807"
   },
   "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>ID</th>\n",
       "      <th>label</th>\n",
       "      <th>sessions</th>\n",
       "      <th>drives</th>\n",
       "      <th>total_sessions</th>\n",
       "      <th>n_days_after_onboarding</th>\n",
       "      <th>total_navigations_fav1</th>\n",
       "      <th>total_navigations_fav2</th>\n",
       "      <th>driven_km_drives</th>\n",
       "      <th>duration_minutes_drives</th>\n",
       "      <th>activity_days</th>\n",
       "      <th>driving_days</th>\n",
       "      <th>device</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>retained</td>\n",
       "      <td>283</td>\n",
       "      <td>226</td>\n",
       "      <td>296.748273</td>\n",
       "      <td>2276</td>\n",
       "      <td>208</td>\n",
       "      <td>0</td>\n",
       "      <td>2628.845068</td>\n",
       "      <td>1985.775061</td>\n",
       "      <td>28</td>\n",
       "      <td>19</td>\n",
       "      <td>Android</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>retained</td>\n",
       "      <td>133</td>\n",
       "      <td>107</td>\n",
       "      <td>326.896596</td>\n",
       "      <td>1225</td>\n",
       "      <td>19</td>\n",
       "      <td>64</td>\n",
       "      <td>13715.920550</td>\n",
       "      <td>3160.472914</td>\n",
       "      <td>13</td>\n",
       "      <td>11</td>\n",
       "      <td>iPhone</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>retained</td>\n",
       "      <td>114</td>\n",
       "      <td>95</td>\n",
       "      <td>135.522926</td>\n",
       "      <td>2651</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3059.148818</td>\n",
       "      <td>1610.735904</td>\n",
       "      <td>14</td>\n",
       "      <td>8</td>\n",
       "      <td>Android</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>retained</td>\n",
       "      <td>49</td>\n",
       "      <td>40</td>\n",
       "      <td>67.589221</td>\n",
       "      <td>15</td>\n",
       "      <td>322</td>\n",
       "      <td>7</td>\n",
       "      <td>913.591123</td>\n",
       "      <td>587.196542</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>iPhone</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>retained</td>\n",
       "      <td>84</td>\n",
       "      <td>68</td>\n",
       "      <td>168.247020</td>\n",
       "      <td>1562</td>\n",
       "      <td>166</td>\n",
       "      <td>5</td>\n",
       "      <td>3950.202008</td>\n",
       "      <td>1219.555924</td>\n",
       "      <td>27</td>\n",
       "      <td>18</td>\n",
       "      <td>Android</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID     label  sessions  drives  total_sessions  n_days_after_onboarding  \\\n",
       "0   0  retained       283     226      296.748273                     2276   \n",
       "1   1  retained       133     107      326.896596                     1225   \n",
       "2   2  retained       114      95      135.522926                     2651   \n",
       "3   3  retained        49      40       67.589221                       15   \n",
       "4   4  retained        84      68      168.247020                     1562   \n",
       "\n",
       "   total_navigations_fav1  total_navigations_fav2  driven_km_drives  \\\n",
       "0                     208                       0       2628.845068   \n",
       "1                      19                      64      13715.920550   \n",
       "2                       0                       0       3059.148818   \n",
       "3                     322                       7        913.591123   \n",
       "4                     166                       5       3950.202008   \n",
       "\n",
       "   duration_minutes_drives  activity_days  driving_days   device  \n",
       "0              1985.775061             28            19  Android  \n",
       "1              3160.472914             13            11   iPhone  \n",
       "2              1610.735904             14             8  Android  \n",
       "3               587.196542              7             3   iPhone  \n",
       "4              1219.555924             27            18  Android  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VSkcBkDhp9P5"
   },
   "source": [
    "Use the `drop()` method to remove the ID column since you don't need this information for your analysis."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "RnFQIV3Cp8Bp"
   },
   "outputs": [],
   "source": [
    "df = df.drop('ID', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8NAFHVG55Y2o"
   },
   "source": [
    "Now, check the class balance of the dependent (target) variable, `label`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 15,
     "status": "ok",
     "timestamp": 1685661316023,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "k5lo-Sz6yizF",
    "outputId": "9dcde8b6-12bb-48fd-a2de-e70af2c46cba"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "retained    0.822645\n",
       "churned     0.177355\n",
       "Name: label, dtype: float64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['label'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pcEIBLGcIbGa"
   },
   "source": [
    "Call `describe()` on the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 364
    },
    "executionInfo": {
     "elapsed": 461,
     "status": "ok",
     "timestamp": 1685661316473,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "AjcSoFeVIhYf",
    "outputId": "c9e52d3d-6ec2-4a00-9bc8-2aa2985bfb6e"
   },
   "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>sessions</th>\n",
       "      <th>drives</th>\n",
       "      <th>total_sessions</th>\n",
       "      <th>n_days_after_onboarding</th>\n",
       "      <th>total_navigations_fav1</th>\n",
       "      <th>total_navigations_fav2</th>\n",
       "      <th>driven_km_drives</th>\n",
       "      <th>duration_minutes_drives</th>\n",
       "      <th>activity_days</th>\n",
       "      <th>driving_days</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "      <td>14999.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>80.633776</td>\n",
       "      <td>67.281152</td>\n",
       "      <td>189.964447</td>\n",
       "      <td>1749.837789</td>\n",
       "      <td>121.605974</td>\n",
       "      <td>29.672512</td>\n",
       "      <td>4039.340921</td>\n",
       "      <td>1860.976012</td>\n",
       "      <td>15.537102</td>\n",
       "      <td>12.179879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>80.699065</td>\n",
       "      <td>65.913872</td>\n",
       "      <td>136.405128</td>\n",
       "      <td>1008.513876</td>\n",
       "      <td>148.121544</td>\n",
       "      <td>45.394651</td>\n",
       "      <td>2502.149334</td>\n",
       "      <td>1446.702288</td>\n",
       "      <td>9.004655</td>\n",
       "      <td>7.824036</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.220211</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>60.441250</td>\n",
       "      <td>18.282082</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>23.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>90.661156</td>\n",
       "      <td>878.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2212.600607</td>\n",
       "      <td>835.996260</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>5.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>56.000000</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>159.568115</td>\n",
       "      <td>1741.000000</td>\n",
       "      <td>71.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>3493.858085</td>\n",
       "      <td>1478.249859</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>12.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>112.000000</td>\n",
       "      <td>93.000000</td>\n",
       "      <td>254.192341</td>\n",
       "      <td>2623.500000</td>\n",
       "      <td>178.000000</td>\n",
       "      <td>43.000000</td>\n",
       "      <td>5289.861262</td>\n",
       "      <td>2464.362632</td>\n",
       "      <td>23.000000</td>\n",
       "      <td>19.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>743.000000</td>\n",
       "      <td>596.000000</td>\n",
       "      <td>1216.154633</td>\n",
       "      <td>3500.000000</td>\n",
       "      <td>1236.000000</td>\n",
       "      <td>415.000000</td>\n",
       "      <td>21183.401890</td>\n",
       "      <td>15851.727160</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>30.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           sessions        drives  total_sessions  n_days_after_onboarding  \\\n",
       "count  14999.000000  14999.000000    14999.000000             14999.000000   \n",
       "mean      80.633776     67.281152      189.964447              1749.837789   \n",
       "std       80.699065     65.913872      136.405128              1008.513876   \n",
       "min        0.000000      0.000000        0.220211                 4.000000   \n",
       "25%       23.000000     20.000000       90.661156               878.000000   \n",
       "50%       56.000000     48.000000      159.568115              1741.000000   \n",
       "75%      112.000000     93.000000      254.192341              2623.500000   \n",
       "max      743.000000    596.000000     1216.154633              3500.000000   \n",
       "\n",
       "       total_navigations_fav1  total_navigations_fav2  driven_km_drives  \\\n",
       "count            14999.000000            14999.000000      14999.000000   \n",
       "mean               121.605974               29.672512       4039.340921   \n",
       "std                148.121544               45.394651       2502.149334   \n",
       "min                  0.000000                0.000000         60.441250   \n",
       "25%                  9.000000                0.000000       2212.600607   \n",
       "50%                 71.000000                9.000000       3493.858085   \n",
       "75%                178.000000               43.000000       5289.861262   \n",
       "max               1236.000000              415.000000      21183.401890   \n",
       "\n",
       "       duration_minutes_drives  activity_days  driving_days  \n",
       "count             14999.000000   14999.000000  14999.000000  \n",
       "mean               1860.976012      15.537102     12.179879  \n",
       "std                1446.702288       9.004655      7.824036  \n",
       "min                  18.282082       0.000000      0.000000  \n",
       "25%                 835.996260       8.000000      5.000000  \n",
       "50%                1478.249859      16.000000     12.000000  \n",
       "75%                2464.362632      23.000000     19.000000  \n",
       "max               15851.727160      31.000000     30.000000  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "BsoAfNJce9r1"
   },
   "source": [
    "**Question:** Are there any variables that could potentially have outliers just by assessing at the quartile values, standard deviation, and max values?\n",
    "\n",
    "> *Yes, the following columns all seem to have outliers:*\n",
    "<br>\n",
    "* `sessions`\n",
    "* `drives`\n",
    "* `total_sessions`\n",
    "* `total_navigations_fav1`\n",
    "* `total_navigations_fav2`\n",
    "* `driven_km_drives`\n",
    "* `duration_minutes_drives`\n",
    "\n",
    "\n",
    "> *All of these columns have max values that are multiple standard deviations above the 75th percentile. This could indicate outliers in these variables.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Xn43i4Gh-Q7M"
   },
   "source": [
    "### **Task 2b. Create features**\n",
    "\n",
    "Create features that may be of interest to the stakeholder and/or that are needed to address the business scenario/problem."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ky5h_Aum3RK1"
   },
   "source": [
    "#### **`km_per_driving_day`**\n",
    "\n",
    "You know from earlier EDA that churn rate correlates with distance driven per driving day in the last month. It might be helpful to engineer a feature that captures this information.\n",
    "\n",
    "1. Create a new column in `df` called `km_per_driving_day`, which represents the mean distance driven per driving day for each user.\n",
    "\n",
    "2. Call the `describe()` method on the new column."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 34,
     "status": "ok",
     "timestamp": 1685661316473,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "IhgtbbCf8l4P",
    "outputId": "3dc9ac42-76fc-4b39-b58b-6a9e8b7a0ea2"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1.499900e+04\n",
       "mean              inf\n",
       "std               NaN\n",
       "min      3.022063e+00\n",
       "25%      1.672804e+02\n",
       "50%      3.231459e+02\n",
       "75%      7.579257e+02\n",
       "max               inf\n",
       "Name: km_per_driving_day, dtype: float64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 1. Create `km_per_driving_day` column\n",
    "df['km_per_driving_day'] = df['driven_km_drives'] / df['driving_days']\n",
    "\n",
    "# 2. Call `describe()` on the new column\n",
    "df['km_per_driving_day'].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "S5XDZpYy85r6"
   },
   "source": [
    "Note that some values are infinite. This is the result of there being values of zero in the `driving_days` column. Pandas imputes a value of infinity in the corresponding rows of the new column because division by zero is undefined.\n",
    "\n",
    "1. Convert these values from infinity to zero. You can use `np.inf` to refer to a value of infinity.\n",
    "\n",
    "2. Call `describe()` on the `km_per_driving_day` column to verify that it worked."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 33,
     "status": "ok",
     "timestamp": 1685661316474,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "xRphAWpW9DK_",
    "outputId": "afcd893c-aefa-4e0f-dcc3-0907573f9d15"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    14999.000000\n",
       "mean       578.963113\n",
       "std       1030.094384\n",
       "min          0.000000\n",
       "25%        136.238895\n",
       "50%        272.889272\n",
       "75%        558.686918\n",
       "max      15420.234110\n",
       "Name: km_per_driving_day, dtype: float64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 1. Convert infinite values to zero\n",
    "df.loc[df['km_per_driving_day']==np.inf, 'km_per_driving_day'] = 0\n",
    "\n",
    "# 2. Confirm that it worked\n",
    "df['km_per_driving_day'].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "YlTioDBX80q3"
   },
   "source": [
    "#### **`professional_driver`**\n",
    "\n",
    "Create a new, binary feature called `professional_driver` that is a 1 for users who had 60 or more drives <u>**and**</u> drove on 15+ days in the last month.\n",
    "\n",
    "**Note:** The objective is to create a new feature that separates professional drivers from other drivers. In this scenario, domain knowledge and intuition are used to determine these deciding thresholds, but ultimately they are arbitrary."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xVBwS_GaHFBe"
   },
   "source": [
    "To create this column, use the [`np.where()`](https://numpy.org/doc/stable/reference/generated/numpy.where.html) function. This function accepts as arguments:\n",
    "1. A condition\n",
    "2. What to return when the condition is true\n",
    "3. What to return when the condition is false\n",
    "\n",
    "```\n",
    "Example:\n",
    "x = [1, 2, 3]\n",
    "x = np.where(x > 2, 100, 0)\n",
    "x\n",
    "array([  0,   0, 100])\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "heHnO3KtDVqu"
   },
   "outputs": [],
   "source": [
    "# Create `professional_driver` column\n",
    "df['professional_driver'] = np.where((df['drives'] >= 60) & (df['driving_days'] >= 15), 1, 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "5BJFC64ULdhS"
   },
   "source": [
    "Perform a quick inspection of the new variable.\n",
    "\n",
    "1. Check the count of professional drivers and non-professionals\n",
    "\n",
    "2. Within each class (professional and non-professional) calculate the churn rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 30,
     "status": "ok",
     "timestamp": 1685661316476,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "Gqz5wCbRJ2-7",
    "outputId": "6e762532-9e7d-49d3-ad12-e8fb32cd0849"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0    12405\n",
      "1     2594\n",
      "Name: professional_driver, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "professional_driver  label   \n",
       "0                    retained    0.801202\n",
       "                     churned     0.198798\n",
       "1                    retained    0.924437\n",
       "                     churned     0.075563\n",
       "Name: label, dtype: float64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 1. Check count of professionals and non-professionals\n",
    "print(df['professional_driver'].value_counts())\n",
    "\n",
    "# 2. Check in-class churn rate\n",
    "df.groupby(['professional_driver'])['label'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LaJxfLvNL97O"
   },
   "source": [
    "The churn rate for professional drivers is 7.6%, while the churn rate for non-professionals is 19.9%. This seems like it could add predictive signal to the model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lgPul2DiY6T4"
   },
   "source": [
    "<img src='images/Construct.png' width='100' height='100' align=left>\n",
    "\n",
    "## **PACE: Construct**\n",
    "\n",
    " After analysis and deriving variables with close relationships, it is time to begin constructing the model. Consider the questions in your PACE Strategy Document to reflect on the Construct stage.\n",
    "\n",
    " In this stage, consider the following question:\n",
    "\n",
    "* Why did you select the X variables you did?\n",
    "\n",
    "> *Initially, columns were dropped based on high multicollinearity. Later, variable selection can be fine-tuned by running and rerunning models to look at changes in accuracy, recall, and precision.*\n",
    "<br></br>\n",
    "> *Initial variable selection was based on the business objective and insights from prior EDA.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "07UJJm41ajgf"
   },
   "source": [
    "### **Task 3a. Preparing variables**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aanTG0VxniQh"
   },
   "source": [
    "Call `info()` on the dataframe to check the data type of the `label` variable and to verify if there are any missing values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 26,
     "status": "ok",
     "timestamp": 1685661316476,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "tHFNCNj3ob30",
    "outputId": "a8eaaa4d-828b-44f2-b47a-3c34395a9ded"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 14999 entries, 0 to 14998\n",
      "Data columns (total 14 columns):\n",
      " #   Column                   Non-Null Count  Dtype  \n",
      "---  ------                   --------------  -----  \n",
      " 0   label                    14299 non-null  object \n",
      " 1   sessions                 14999 non-null  int64  \n",
      " 2   drives                   14999 non-null  int64  \n",
      " 3   total_sessions           14999 non-null  float64\n",
      " 4   n_days_after_onboarding  14999 non-null  int64  \n",
      " 5   total_navigations_fav1   14999 non-null  int64  \n",
      " 6   total_navigations_fav2   14999 non-null  int64  \n",
      " 7   driven_km_drives         14999 non-null  float64\n",
      " 8   duration_minutes_drives  14999 non-null  float64\n",
      " 9   activity_days            14999 non-null  int64  \n",
      " 10  driving_days             14999 non-null  int64  \n",
      " 11  device                   14999 non-null  object \n",
      " 12  km_per_driving_day       14999 non-null  float64\n",
      " 13  professional_driver      14999 non-null  int64  \n",
      "dtypes: float64(4), int64(8), object(2)\n",
      "memory usage: 1.6+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "J90QgkTrofOC"
   },
   "source": [
    "Because you know from previous EDA that there is no evidence of a non-random cause of the 700 missing values in the `label` column, and because these observations comprise less than 5% of the data, use the `dropna()` method to drop the rows that are missing this data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "B34fDk52o2Uk"
   },
   "outputs": [],
   "source": [
    "# Drop rows with missing data in `label` column\n",
    "df = df.dropna(subset=['label'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TrmIiSQ-PG8Q"
   },
   "source": [
    "#### **Impute outliers**\n",
    "\n",
    "You rarely want to drop outliers, and generally will not do so unless there is a clear reason for it (e.g., typographic errors).\n",
    "\n",
    "At times outliers can be changed to the **median, mean, 95th percentile, etc.**\n",
    "\n",
    "Previously, you determined that seven of the variables had clear signs of containing outliers:\n",
    "\n",
    "* `sessions`\n",
    "* `drives`\n",
    "* `total_sessions`\n",
    "* `total_navigations_fav1`\n",
    "* `total_navigations_fav2`\n",
    "* `driven_km_drives`\n",
    "* `duration_minutes_drives`\n",
    "\n",
    "For this analysis, impute the outlying values for these columns. Calculate the **95th percentile** of each column and change to this value any value in the column that exceeds it.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "id": "AX-Sbttr9-ji"
   },
   "outputs": [],
   "source": [
    "# Impute outliers\n",
    "for column in ['sessions', 'drives', 'total_sessions', 'total_navigations_fav1',\n",
    "               'total_navigations_fav2', 'driven_km_drives', 'duration_minutes_drives']:\n",
    "    threshold = df[column].quantile(0.95)\n",
    "    df.loc[df[column] > threshold, column] = threshold"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ih9_NxMrAihX"
   },
   "source": [
    "Call `describe()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 24,
     "status": "ok",
     "timestamp": 1685661316480,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "isYOM__qS75C",
    "outputId": "6bef6f00-946e-4a05-ab0b-2948ae28a907"
   },
   "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>sessions</th>\n",
       "      <th>drives</th>\n",
       "      <th>total_sessions</th>\n",
       "      <th>n_days_after_onboarding</th>\n",
       "      <th>total_navigations_fav1</th>\n",
       "      <th>total_navigations_fav2</th>\n",
       "      <th>driven_km_drives</th>\n",
       "      <th>duration_minutes_drives</th>\n",
       "      <th>activity_days</th>\n",
       "      <th>driving_days</th>\n",
       "      <th>km_per_driving_day</th>\n",
       "      <th>professional_driver</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "      <td>14299.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>76.539688</td>\n",
       "      <td>63.964683</td>\n",
       "      <td>183.717304</td>\n",
       "      <td>1751.822505</td>\n",
       "      <td>114.562767</td>\n",
       "      <td>27.187216</td>\n",
       "      <td>3944.558631</td>\n",
       "      <td>1792.911210</td>\n",
       "      <td>15.544653</td>\n",
       "      <td>12.182530</td>\n",
       "      <td>581.942399</td>\n",
       "      <td>0.173998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>67.243178</td>\n",
       "      <td>55.127927</td>\n",
       "      <td>118.720520</td>\n",
       "      <td>1008.663834</td>\n",
       "      <td>124.378550</td>\n",
       "      <td>36.715302</td>\n",
       "      <td>2218.358258</td>\n",
       "      <td>1224.329759</td>\n",
       "      <td>9.016088</td>\n",
       "      <td>7.833835</td>\n",
       "      <td>1038.254509</td>\n",
       "      <td>0.379121</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.220211</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>60.441250</td>\n",
       "      <td>18.282082</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>23.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>90.457733</td>\n",
       "      <td>878.500000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2217.319909</td>\n",
       "      <td>840.181344</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>136.168003</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>56.000000</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>158.718571</td>\n",
       "      <td>1749.000000</td>\n",
       "      <td>71.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>3496.545617</td>\n",
       "      <td>1479.394387</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>273.301012</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>111.000000</td>\n",
       "      <td>93.000000</td>\n",
       "      <td>253.540450</td>\n",
       "      <td>2627.500000</td>\n",
       "      <td>178.000000</td>\n",
       "      <td>43.000000</td>\n",
       "      <td>5299.972162</td>\n",
       "      <td>2466.928876</td>\n",
       "      <td>23.000000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>558.018761</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>243.000000</td>\n",
       "      <td>200.000000</td>\n",
       "      <td>455.439492</td>\n",
       "      <td>3500.000000</td>\n",
       "      <td>422.000000</td>\n",
       "      <td>124.000000</td>\n",
       "      <td>8898.716275</td>\n",
       "      <td>4668.180092</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>15420.234110</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           sessions        drives  total_sessions  n_days_after_onboarding  \\\n",
       "count  14299.000000  14299.000000    14299.000000             14299.000000   \n",
       "mean      76.539688     63.964683      183.717304              1751.822505   \n",
       "std       67.243178     55.127927      118.720520              1008.663834   \n",
       "min        0.000000      0.000000        0.220211                 4.000000   \n",
       "25%       23.000000     20.000000       90.457733               878.500000   \n",
       "50%       56.000000     48.000000      158.718571              1749.000000   \n",
       "75%      111.000000     93.000000      253.540450              2627.500000   \n",
       "max      243.000000    200.000000      455.439492              3500.000000   \n",
       "\n",
       "       total_navigations_fav1  total_navigations_fav2  driven_km_drives  \\\n",
       "count            14299.000000            14299.000000      14299.000000   \n",
       "mean               114.562767               27.187216       3944.558631   \n",
       "std                124.378550               36.715302       2218.358258   \n",
       "min                  0.000000                0.000000         60.441250   \n",
       "25%                 10.000000                0.000000       2217.319909   \n",
       "50%                 71.000000                9.000000       3496.545617   \n",
       "75%                178.000000               43.000000       5299.972162   \n",
       "max                422.000000              124.000000       8898.716275   \n",
       "\n",
       "       duration_minutes_drives  activity_days  driving_days  \\\n",
       "count             14299.000000   14299.000000  14299.000000   \n",
       "mean               1792.911210      15.544653     12.182530   \n",
       "std                1224.329759       9.016088      7.833835   \n",
       "min                  18.282082       0.000000      0.000000   \n",
       "25%                 840.181344       8.000000      5.000000   \n",
       "50%                1479.394387      16.000000     12.000000   \n",
       "75%                2466.928876      23.000000     19.000000   \n",
       "max                4668.180092      31.000000     30.000000   \n",
       "\n",
       "       km_per_driving_day  professional_driver  \n",
       "count        14299.000000         14299.000000  \n",
       "mean           581.942399             0.173998  \n",
       "std           1038.254509             0.379121  \n",
       "min              0.000000             0.000000  \n",
       "25%            136.168003             0.000000  \n",
       "50%            273.301012             0.000000  \n",
       "75%            558.018761             0.000000  \n",
       "max          15420.234110             1.000000  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uFnVVQ_OTpqz"
   },
   "source": [
    "#### **Encode categorical variables**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "UV3KLq7LpcWB"
   },
   "source": [
    "Change the data type of the `label` column to be binary. This change is needed to train a logistic regression model.\n",
    "\n",
    "Assign a `0` for all `retained` users.\n",
    "\n",
    "Assign a `1` for all `churned` users.\n",
    "\n",
    "Save this variable as `label2` as to not overwrite the original `label` variable.\n",
    "\n",
    "**Note:** There are many ways to do this. Consider using `np.where()` as you did earlier in this notebook.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "executionInfo": {
     "elapsed": 19,
     "status": "ok",
     "timestamp": 1685661316480,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "7fRaU2JKpyXg",
    "outputId": "2d50426f-bf38-48e2-8b07-28825c2826bd"
   },
   "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>label</th>\n",
       "      <th>label2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14994</th>\n",
       "      <td>retained</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14995</th>\n",
       "      <td>retained</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14996</th>\n",
       "      <td>retained</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14997</th>\n",
       "      <td>churned</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14998</th>\n",
       "      <td>retained</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          label  label2\n",
       "14994  retained       0\n",
       "14995  retained       0\n",
       "14996  retained       0\n",
       "14997   churned       1\n",
       "14998  retained       0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create binary `label2` column\n",
    "df['label2'] = np.where(df['label']=='churned', 1, 0)\n",
    "df[['label', 'label2']].tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HMj6QkK1cLmS"
   },
   "source": [
    "### **Task 3b. Determine whether assumptions have been met**\n",
    "\n",
    "The following are the assumptions for logistic regression:\n",
    "\n",
    "* Independent observations (This refers to how the data was collected.)\n",
    "\n",
    "* No extreme outliers\n",
    "\n",
    "* Little to no multicollinearity among X predictors\n",
    "\n",
    "* Linear relationship between X and the **logit** of y\n",
    "\n",
    "For the first assumption, you can assume that observations are independent for this project.\n",
    "\n",
    "The second assumption has already been addressed.\n",
    "\n",
    "The last assumption will be verified after modeling.\n",
    "\n",
    "**Note:** In practice, modeling assumptions are often violated, and depending on the specifics of your use case and the severity of the violation, it might not affect your model much at all or it will result in a failed model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "zl5_1hXIWoqw"
   },
   "source": [
    "#### **Collinearity**\n",
    "\n",
    "Check the correlation among predictor variables. First, generate a correlation matrix."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 521
    },
    "executionInfo": {
     "elapsed": 19,
     "status": "ok",
     "timestamp": 1685661316481,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "10gLfs4eW8wl",
    "outputId": "6c36a069-a532-42a7-ef21-7946caaa53da"
   },
   "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>sessions</th>\n",
       "      <th>drives</th>\n",
       "      <th>total_sessions</th>\n",
       "      <th>n_days_after_onboarding</th>\n",
       "      <th>total_navigations_fav1</th>\n",
       "      <th>total_navigations_fav2</th>\n",
       "      <th>driven_km_drives</th>\n",
       "      <th>duration_minutes_drives</th>\n",
       "      <th>activity_days</th>\n",
       "      <th>driving_days</th>\n",
       "      <th>km_per_driving_day</th>\n",
       "      <th>professional_driver</th>\n",
       "      <th>label2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>sessions</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.996942</td>\n",
       "      <td>0.597189</td>\n",
       "      <td>0.007101</td>\n",
       "      <td>0.001858</td>\n",
       "      <td>0.008536</td>\n",
       "      <td>0.002996</td>\n",
       "      <td>-0.004545</td>\n",
       "      <td>0.025113</td>\n",
       "      <td>0.020294</td>\n",
       "      <td>-0.011569</td>\n",
       "      <td>0.443654</td>\n",
       "      <td>0.034911</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>drives</th>\n",
       "      <td>0.996942</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.595285</td>\n",
       "      <td>0.006940</td>\n",
       "      <td>0.001058</td>\n",
       "      <td>0.009505</td>\n",
       "      <td>0.003445</td>\n",
       "      <td>-0.003889</td>\n",
       "      <td>0.024357</td>\n",
       "      <td>0.019608</td>\n",
       "      <td>-0.010989</td>\n",
       "      <td>0.444425</td>\n",
       "      <td>0.035865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>total_sessions</th>\n",
       "      <td>0.597189</td>\n",
       "      <td>0.595285</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.006596</td>\n",
       "      <td>0.000187</td>\n",
       "      <td>0.010371</td>\n",
       "      <td>0.001016</td>\n",
       "      <td>-0.000338</td>\n",
       "      <td>0.015755</td>\n",
       "      <td>0.012953</td>\n",
       "      <td>-0.016167</td>\n",
       "      <td>0.254433</td>\n",
       "      <td>0.024568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>n_days_after_onboarding</th>\n",
       "      <td>0.007101</td>\n",
       "      <td>0.006940</td>\n",
       "      <td>0.006596</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.002450</td>\n",
       "      <td>-0.004968</td>\n",
       "      <td>-0.004652</td>\n",
       "      <td>-0.010167</td>\n",
       "      <td>-0.009418</td>\n",
       "      <td>-0.007321</td>\n",
       "      <td>0.011764</td>\n",
       "      <td>0.003770</td>\n",
       "      <td>-0.129263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>total_navigations_fav1</th>\n",
       "      <td>0.001858</td>\n",
       "      <td>0.001058</td>\n",
       "      <td>0.000187</td>\n",
       "      <td>-0.002450</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.002866</td>\n",
       "      <td>-0.007368</td>\n",
       "      <td>0.005646</td>\n",
       "      <td>0.010902</td>\n",
       "      <td>0.010419</td>\n",
       "      <td>-0.000197</td>\n",
       "      <td>-0.000224</td>\n",
       "      <td>0.052322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>total_navigations_fav2</th>\n",
       "      <td>0.008536</td>\n",
       "      <td>0.009505</td>\n",
       "      <td>0.010371</td>\n",
       "      <td>-0.004968</td>\n",
       "      <td>0.002866</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.003559</td>\n",
       "      <td>-0.003009</td>\n",
       "      <td>-0.004425</td>\n",
       "      <td>0.002000</td>\n",
       "      <td>0.006751</td>\n",
       "      <td>0.007126</td>\n",
       "      <td>0.015032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>driven_km_drives</th>\n",
       "      <td>0.002996</td>\n",
       "      <td>0.003445</td>\n",
       "      <td>0.001016</td>\n",
       "      <td>-0.004652</td>\n",
       "      <td>-0.007368</td>\n",
       "      <td>0.003559</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.690515</td>\n",
       "      <td>-0.007441</td>\n",
       "      <td>-0.009549</td>\n",
       "      <td>0.344811</td>\n",
       "      <td>-0.000904</td>\n",
       "      <td>0.019767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>duration_minutes_drives</th>\n",
       "      <td>-0.004545</td>\n",
       "      <td>-0.003889</td>\n",
       "      <td>-0.000338</td>\n",
       "      <td>-0.010167</td>\n",
       "      <td>0.005646</td>\n",
       "      <td>-0.003009</td>\n",
       "      <td>0.690515</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.007895</td>\n",
       "      <td>-0.009425</td>\n",
       "      <td>0.239627</td>\n",
       "      <td>-0.012128</td>\n",
       "      <td>0.040407</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>activity_days</th>\n",
       "      <td>0.025113</td>\n",
       "      <td>0.024357</td>\n",
       "      <td>0.015755</td>\n",
       "      <td>-0.009418</td>\n",
       "      <td>0.010902</td>\n",
       "      <td>-0.004425</td>\n",
       "      <td>-0.007441</td>\n",
       "      <td>-0.007895</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.947687</td>\n",
       "      <td>-0.397433</td>\n",
       "      <td>0.453825</td>\n",
       "      <td>-0.303851</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>driving_days</th>\n",
       "      <td>0.020294</td>\n",
       "      <td>0.019608</td>\n",
       "      <td>0.012953</td>\n",
       "      <td>-0.007321</td>\n",
       "      <td>0.010419</td>\n",
       "      <td>0.002000</td>\n",
       "      <td>-0.009549</td>\n",
       "      <td>-0.009425</td>\n",
       "      <td>0.947687</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.407917</td>\n",
       "      <td>0.469776</td>\n",
       "      <td>-0.294259</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>km_per_driving_day</th>\n",
       "      <td>-0.011569</td>\n",
       "      <td>-0.010989</td>\n",
       "      <td>-0.016167</td>\n",
       "      <td>0.011764</td>\n",
       "      <td>-0.000197</td>\n",
       "      <td>0.006751</td>\n",
       "      <td>0.344811</td>\n",
       "      <td>0.239627</td>\n",
       "      <td>-0.397433</td>\n",
       "      <td>-0.407917</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.165966</td>\n",
       "      <td>0.148583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>professional_driver</th>\n",
       "      <td>0.443654</td>\n",
       "      <td>0.444425</td>\n",
       "      <td>0.254433</td>\n",
       "      <td>0.003770</td>\n",
       "      <td>-0.000224</td>\n",
       "      <td>0.007126</td>\n",
       "      <td>-0.000904</td>\n",
       "      <td>-0.012128</td>\n",
       "      <td>0.453825</td>\n",
       "      <td>0.469776</td>\n",
       "      <td>-0.165966</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.122312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>label2</th>\n",
       "      <td>0.034911</td>\n",
       "      <td>0.035865</td>\n",
       "      <td>0.024568</td>\n",
       "      <td>-0.129263</td>\n",
       "      <td>0.052322</td>\n",
       "      <td>0.015032</td>\n",
       "      <td>0.019767</td>\n",
       "      <td>0.040407</td>\n",
       "      <td>-0.303851</td>\n",
       "      <td>-0.294259</td>\n",
       "      <td>0.148583</td>\n",
       "      <td>-0.122312</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         sessions    drives  total_sessions  \\\n",
       "sessions                 1.000000  0.996942        0.597189   \n",
       "drives                   0.996942  1.000000        0.595285   \n",
       "total_sessions           0.597189  0.595285        1.000000   \n",
       "n_days_after_onboarding  0.007101  0.006940        0.006596   \n",
       "total_navigations_fav1   0.001858  0.001058        0.000187   \n",
       "total_navigations_fav2   0.008536  0.009505        0.010371   \n",
       "driven_km_drives         0.002996  0.003445        0.001016   \n",
       "duration_minutes_drives -0.004545 -0.003889       -0.000338   \n",
       "activity_days            0.025113  0.024357        0.015755   \n",
       "driving_days             0.020294  0.019608        0.012953   \n",
       "km_per_driving_day      -0.011569 -0.010989       -0.016167   \n",
       "professional_driver      0.443654  0.444425        0.254433   \n",
       "label2                   0.034911  0.035865        0.024568   \n",
       "\n",
       "                         n_days_after_onboarding  total_navigations_fav1  \\\n",
       "sessions                                0.007101                0.001858   \n",
       "drives                                  0.006940                0.001058   \n",
       "total_sessions                          0.006596                0.000187   \n",
       "n_days_after_onboarding                 1.000000               -0.002450   \n",
       "total_navigations_fav1                 -0.002450                1.000000   \n",
       "total_navigations_fav2                 -0.004968                0.002866   \n",
       "driven_km_drives                       -0.004652               -0.007368   \n",
       "duration_minutes_drives                -0.010167                0.005646   \n",
       "activity_days                          -0.009418                0.010902   \n",
       "driving_days                           -0.007321                0.010419   \n",
       "km_per_driving_day                      0.011764               -0.000197   \n",
       "professional_driver                     0.003770               -0.000224   \n",
       "label2                                 -0.129263                0.052322   \n",
       "\n",
       "                         total_navigations_fav2  driven_km_drives  \\\n",
       "sessions                               0.008536          0.002996   \n",
       "drives                                 0.009505          0.003445   \n",
       "total_sessions                         0.010371          0.001016   \n",
       "n_days_after_onboarding               -0.004968         -0.004652   \n",
       "total_navigations_fav1                 0.002866         -0.007368   \n",
       "total_navigations_fav2                 1.000000          0.003559   \n",
       "driven_km_drives                       0.003559          1.000000   \n",
       "duration_minutes_drives               -0.003009          0.690515   \n",
       "activity_days                         -0.004425         -0.007441   \n",
       "driving_days                           0.002000         -0.009549   \n",
       "km_per_driving_day                     0.006751          0.344811   \n",
       "professional_driver                    0.007126         -0.000904   \n",
       "label2                                 0.015032          0.019767   \n",
       "\n",
       "                         duration_minutes_drives  activity_days  driving_days  \\\n",
       "sessions                               -0.004545       0.025113      0.020294   \n",
       "drives                                 -0.003889       0.024357      0.019608   \n",
       "total_sessions                         -0.000338       0.015755      0.012953   \n",
       "n_days_after_onboarding                -0.010167      -0.009418     -0.007321   \n",
       "total_navigations_fav1                  0.005646       0.010902      0.010419   \n",
       "total_navigations_fav2                 -0.003009      -0.004425      0.002000   \n",
       "driven_km_drives                        0.690515      -0.007441     -0.009549   \n",
       "duration_minutes_drives                 1.000000      -0.007895     -0.009425   \n",
       "activity_days                          -0.007895       1.000000      0.947687   \n",
       "driving_days                           -0.009425       0.947687      1.000000   \n",
       "km_per_driving_day                      0.239627      -0.397433     -0.407917   \n",
       "professional_driver                    -0.012128       0.453825      0.469776   \n",
       "label2                                  0.040407      -0.303851     -0.294259   \n",
       "\n",
       "                         km_per_driving_day  professional_driver    label2  \n",
       "sessions                          -0.011569             0.443654  0.034911  \n",
       "drives                            -0.010989             0.444425  0.035865  \n",
       "total_sessions                    -0.016167             0.254433  0.024568  \n",
       "n_days_after_onboarding            0.011764             0.003770 -0.129263  \n",
       "total_navigations_fav1            -0.000197            -0.000224  0.052322  \n",
       "total_navigations_fav2             0.006751             0.007126  0.015032  \n",
       "driven_km_drives                   0.344811            -0.000904  0.019767  \n",
       "duration_minutes_drives            0.239627            -0.012128  0.040407  \n",
       "activity_days                     -0.397433             0.453825 -0.303851  \n",
       "driving_days                      -0.407917             0.469776 -0.294259  \n",
       "km_per_driving_day                 1.000000            -0.165966  0.148583  \n",
       "professional_driver               -0.165966             1.000000 -0.122312  \n",
       "label2                             0.148583            -0.122312  1.000000  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Generate a correlation matrix\n",
    "df.corr(method='pearson')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Zc7o9zzzXDPr"
   },
   "source": [
    "Now, plot a correlation heatmap."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 800
    },
    "executionInfo": {
     "elapsed": 977,
     "status": "ok",
     "timestamp": 1685661317441,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "KLm9uCkpXITf",
    "outputId": "073ae2e1-5ffe-4627-a586-0e9a730addc1"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot correlation heatmap\n",
    "plt.figure(figsize=(15,10))\n",
    "sns.heatmap(df.corr(method='pearson'), vmin=-1, vmax=1, annot=True, cmap='coolwarm')\n",
    "plt.title('Correlation heatmap indicates many low correlated variables',\n",
    "          fontsize=18)\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Lma-Rk81XbA2"
   },
   "source": [
    "If there are predictor variables that have a Pearson correlation coefficient value greater than the **absolute value of 0.7**, these variables are strongly multicollinear. Therefore, only one of these variables should be used in your model.\n",
    "\n",
    "**Note:** 0.7 is an arbitrary threshold. Some industries may use 0.6, 0.8, etc.\n",
    "\n",
    "**Question:** Which variables are multicollinear with each other?\n",
    "\n",
    "> * *`sessions` and `drives`: 1.0*\n",
    "<br>\n",
    "> * *`driving_days` and `activity_days`: 0.95*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "o3ArC_5xa7Oi"
   },
   "source": [
    "### **Task 3c. Create dummies (if necessary)**\n",
    "\n",
    "If you have selected `device` as an X variable, you will need to create dummy variables since this variable is categorical.\n",
    "\n",
    "In cases with many categorical variables, you can use pandas built-in [`pd.get_dummies()`](https://pandas.pydata.org/docs/reference/api/pandas.get_dummies.html), or you can use scikit-learn's [`OneHotEncoder()`](https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.OneHotEncoder.html) function.\n",
    "\n",
    "**Note:** Variables with many categories should only be dummied if absolutely necessary. Each category will result in a coefficient for your model which can lead to overfitting.\n",
    "\n",
    "Because this dataset only has one remaining categorical feature (`device`), it's not necessary to use one of these special functions. You can just implement the transformation directly.\n",
    "\n",
    "Create a new, binary column called `device2` that encodes user devices as follows:\n",
    "\n",
    "* `Android` -> `0`\n",
    "* `iPhone` -> `1`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "executionInfo": {
     "elapsed": 35,
     "status": "ok",
     "timestamp": 1685661317442,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "QvDpwcQm0f35",
    "outputId": "59a2ada0-b47c-43dd-d570-156d042f9e4a"
   },
   "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>device</th>\n",
       "      <th>device2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14994</th>\n",
       "      <td>iPhone</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14995</th>\n",
       "      <td>Android</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14996</th>\n",
       "      <td>iPhone</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14997</th>\n",
       "      <td>iPhone</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14998</th>\n",
       "      <td>iPhone</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        device  device2\n",
       "14994   iPhone        1\n",
       "14995  Android        0\n",
       "14996   iPhone        1\n",
       "14997   iPhone        1\n",
       "14998   iPhone        1"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create new `device2` variable\n",
    "df['device2'] = np.where(df['device']=='Android', 0, 1)\n",
    "df[['device', 'device2']].tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "iDYyjWssbnBG"
   },
   "source": [
    "### **Task 3d. Model building**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VdMw1FZIixYG"
   },
   "source": [
    "#### **Assign predictor variables and target**\n",
    "\n",
    "To build your model you need to determine what X variables you want to include in your model to predict your target&mdash;`label2`.\n",
    "\n",
    "Drop the following variables and assign the results to `X`:\n",
    "\n",
    "* `label` (this is the target)\n",
    "* `label2` (this is the target)\n",
    "* `device` (this is the non-binary-encoded categorical variable)\n",
    "* `sessions` (this had high multicollinearity)\n",
    "* `driving_days` (this had high multicollinearity)\n",
    "\n",
    "**Note:** Notice that `sessions` and `driving_days` were selected to be dropped, rather than `drives` and `activity_days`. The reason for this is that the features that were kept for modeling had slightly stronger correlations with the target variable than the features that were dropped."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "AzcDgLRET4d7"
   },
   "outputs": [],
   "source": [
    "# Isolate predictor variables\n",
    "X = df.drop(columns = ['label', 'label2', 'device', 'sessions', 'driving_days'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Fp1JXgNpigF7"
   },
   "source": [
    "Now, isolate the dependent (target) variable. Assign it to a variable called `y`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "id": "VinAdTeDipKP"
   },
   "outputs": [],
   "source": [
    "# Isolate target variable\n",
    "y = df['label2']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VOewKY740Beq"
   },
   "source": [
    "#### **Split the data**\n",
    "\n",
    "Use scikit-learn's [`train_test_split()`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html) function to perform a train/test split on your data using the X and y variables you assigned above.\n",
    "\n",
    "**Note 1:** It is important to do a train test to obtain accurate predictions.  You always want to fit your model on your training set and evaluate your model on your test set to avoid data leakage.\n",
    "\n",
    "**Note 2:** Because the target class is imbalanced (82% retained vs. 18% churned), you want to make sure that you don't get an unlucky split that over- or under-represents the frequency of the minority class. Set the function's `stratify` parameter to `y` to ensure that the minority class appears in both train and test sets in the same proportion that it does in the overall dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "id": "ulDZdfSS0FyD"
   },
   "outputs": [],
   "source": [
    "# Perform the train-test split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 270
    },
    "executionInfo": {
     "elapsed": 34,
     "status": "ok",
     "timestamp": 1685661317445,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "gdCOXqi4nqwM",
    "outputId": "f789bc02-cbaf-4866-999e-d55df2905ce4"
   },
   "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>drives</th>\n",
       "      <th>total_sessions</th>\n",
       "      <th>n_days_after_onboarding</th>\n",
       "      <th>total_navigations_fav1</th>\n",
       "      <th>total_navigations_fav2</th>\n",
       "      <th>driven_km_drives</th>\n",
       "      <th>duration_minutes_drives</th>\n",
       "      <th>activity_days</th>\n",
       "      <th>km_per_driving_day</th>\n",
       "      <th>professional_driver</th>\n",
       "      <th>device2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>152</th>\n",
       "      <td>108</td>\n",
       "      <td>186.192746</td>\n",
       "      <td>3116</td>\n",
       "      <td>243</td>\n",
       "      <td>124</td>\n",
       "      <td>8898.716275</td>\n",
       "      <td>4668.180092</td>\n",
       "      <td>24</td>\n",
       "      <td>612.305861</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11899</th>\n",
       "      <td>2</td>\n",
       "      <td>3.487590</td>\n",
       "      <td>794</td>\n",
       "      <td>114</td>\n",
       "      <td>18</td>\n",
       "      <td>3286.545691</td>\n",
       "      <td>1780.902733</td>\n",
       "      <td>5</td>\n",
       "      <td>3286.545691</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10937</th>\n",
       "      <td>139</td>\n",
       "      <td>347.106403</td>\n",
       "      <td>331</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>7400.838975</td>\n",
       "      <td>2349.305267</td>\n",
       "      <td>15</td>\n",
       "      <td>616.736581</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>669</th>\n",
       "      <td>108</td>\n",
       "      <td>455.439492</td>\n",
       "      <td>2320</td>\n",
       "      <td>11</td>\n",
       "      <td>4</td>\n",
       "      <td>6566.424830</td>\n",
       "      <td>4558.459870</td>\n",
       "      <td>18</td>\n",
       "      <td>410.401552</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8406</th>\n",
       "      <td>10</td>\n",
       "      <td>89.475821</td>\n",
       "      <td>2478</td>\n",
       "      <td>135</td>\n",
       "      <td>0</td>\n",
       "      <td>1271.248661</td>\n",
       "      <td>938.711572</td>\n",
       "      <td>27</td>\n",
       "      <td>74.779333</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       drives  total_sessions  n_days_after_onboarding  \\\n",
       "152       108      186.192746                     3116   \n",
       "11899       2        3.487590                      794   \n",
       "10937     139      347.106403                      331   \n",
       "669       108      455.439492                     2320   \n",
       "8406       10       89.475821                     2478   \n",
       "\n",
       "       total_navigations_fav1  total_navigations_fav2  driven_km_drives  \\\n",
       "152                       243                     124       8898.716275   \n",
       "11899                     114                      18       3286.545691   \n",
       "10937                       4                       7       7400.838975   \n",
       "669                        11                       4       6566.424830   \n",
       "8406                      135                       0       1271.248661   \n",
       "\n",
       "       duration_minutes_drives  activity_days  km_per_driving_day  \\\n",
       "152                4668.180092             24          612.305861   \n",
       "11899              1780.902733              5         3286.545691   \n",
       "10937              2349.305267             15          616.736581   \n",
       "669                4558.459870             18          410.401552   \n",
       "8406                938.711572             27           74.779333   \n",
       "\n",
       "       professional_driver  device2  \n",
       "152                      1        1  \n",
       "11899                    0        1  \n",
       "10937                    0        0  \n",
       "669                      1        1  \n",
       "8406                     0        1  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use .head()\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ZMUo8Ri-zK7r"
   },
   "source": [
    "Use scikit-learn to instantiate a logistic regression model. Add the argument `penalty = None`.\n",
    "\n",
    "It is important to add `penalty = 'none'` since your predictors are unscaled.\n",
    "\n",
    "Refer to scikit-learn's [logistic regression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) documentation for more information.\n",
    "\n",
    "Fit the model on `X_train` and `y_train`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 74
    },
    "executionInfo": {
     "elapsed": 33,
     "status": "ok",
     "timestamp": 1685661317446,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "zrTNaDVZheyp",
    "outputId": "2dd7cf30-2723-4667-aab3-ecca0a757c05"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=400,\n",
       "                   multi_class='auto', n_jobs=None, penalty='none',\n",
       "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = LogisticRegression(penalty='none', max_iter=400)\n",
    "\n",
    "model.fit(X_train, y_train)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aPoDllWB6ULV"
   },
   "source": [
    "Call the `.coef_` attribute on the model to get the coefficients of each variable.  The coefficients are in order of how the variables are listed in the dataset.  Remember that the coefficients represent the change in the **log odds** of the target variable for **every one unit increase in X**.\n",
    "\n",
    "If you want, create a series whose index is the column names and whose values are the coefficients in `model.coef_`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 32,
     "status": "ok",
     "timestamp": 1685661317446,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "7Ri-OHrlmd8j",
    "outputId": "f7c2b295-9194-47a7-b64b-8e40547414c7"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "drives                     0.001913\n",
       "total_sessions             0.000327\n",
       "n_days_after_onboarding   -0.000406\n",
       "total_navigations_fav1     0.001232\n",
       "total_navigations_fav2     0.000931\n",
       "driven_km_drives          -0.000015\n",
       "duration_minutes_drives    0.000109\n",
       "activity_days             -0.106032\n",
       "km_per_driving_day         0.000018\n",
       "professional_driver       -0.001529\n",
       "device2                   -0.001041\n",
       "dtype: float64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(model.coef_[0], index=X.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_ENmLXXp8JDM"
   },
   "source": [
    "Call the model's `intercept_` attribute to get the intercept of the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 27,
     "status": "ok",
     "timestamp": 1685661317447,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "hWeruvy1wksj",
    "outputId": "8c4599d9-ea2b-443d-a7c3-bc445bb8028e"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.00170675])"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.intercept_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "kGw8Tv7naFIl"
   },
   "source": [
    "#### **Check final assumption**\n",
    "\n",
    "Verify the linear relationship between X and the estimated log odds (known as logits) by making a regplot.\n",
    "\n",
    "Call the model's `predict_proba()` method to generate the probability of response for each sample in the training data. (The training data is the argument to the method.) Assign the result to a variable called `training_probabilities`. This results in a 2-D array where each row represents a user in `X_train`. The first column is the probability of the user not churning, and the second column is the probability of the user churning."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 23,
     "status": "ok",
     "timestamp": 1685661317447,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "GQrFrGIyHCjA",
    "outputId": "c2ea3168-ddfb-4375-abfd-7b3829084933"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.93963483, 0.06036517],\n",
       "       [0.61967304, 0.38032696],\n",
       "       [0.76463181, 0.23536819],\n",
       "       ...,\n",
       "       [0.91909641, 0.08090359],\n",
       "       [0.85092112, 0.14907888],\n",
       "       [0.93516293, 0.06483707]])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get the predicted probabilities of the training data\n",
    "training_probabilities = model.predict_proba(X_train)\n",
    "training_probabilities"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rMET0N5VKALw"
   },
   "source": [
    "In logistic regression, the relationship between a predictor variable and the dependent variable does not need to be linear, however, the log-odds (a.k.a., logit) of the dependent variable with respect to the predictor variable should be linear. Here is the formula for calculating log-odds, where _p_ is the probability of response:\n",
    "<br>\n",
    "$$\n",
    "logit(p) = ln(\\frac{p}{1-p})\n",
    "$$\n",
    "<br>\n",
    "\n",
    "1. Create a dataframe called `logit_data` that is a copy of `df`.\n",
    "\n",
    "2. Create a new column called `logit` in the `logit_data` dataframe. The data in this column should represent the logit for each user.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "id": "7SxkJyInYNCi"
   },
   "outputs": [],
   "source": [
    "# 1. Copy the `X_train` dataframe and assign to `logit_data`\n",
    "logit_data = X_train.copy()\n",
    "\n",
    "# 2. Create a new `logit` column in the `logit_data` df\n",
    "logit_data['logit'] = [np.log(prob[1] / prob[0]) for prob in training_probabilities]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "nlSPpwU1a4kw"
   },
   "source": [
    "Plot a regplot where the x-axis represents an independent variable and the y-axis represents the log-odds of the predicted probabilities.\n",
    "\n",
    "In an exhaustive analysis, this would be plotted for each continuous or discrete predictor variable. Here we show only `activity_days`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 472
    },
    "executionInfo": {
     "elapsed": 809,
     "status": "ok",
     "timestamp": 1685661318647,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "d46KmI_ZZBDy",
    "outputId": "24b6fc58-2323-4071-f6a5-cd6569e3e47a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEXCAYAAACkpJNEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOydd3hc1Zn/v2f6jDQqo2ZZzZJtAa64yCUBG2wTmgkQEkwwZDcklN3Nkk3ZJIRklxQgu6SShJCQH2ETIGBCIOBQgg3GNuDeLXCTrW51jUaj6XN+f9wzQlfonnNkjSRbcz7Po+eV5h7dezWSzrnnLd+XUEqhUCgUitTDNN43oFAoFIrxQS0ACoVCkaKoBUChUChSFLUAKBQKRYqiFgCFQqFIUdQCoFAoFCmKWgAU5zyEkPsIIU9yjp8ihKway3viQQj5NiHk95JjewkhFaN8P9z3TzFxUQuAYlicbZPp2Q4h5BJCSMPA1yilD1BKvyjz/ZTSdEppDTvXE4SQH47GfSpSE7UAKBQKRYqiFgBFUiCE2AkhPyeENLGPnxNC7AOOf4MQ0syOfZEQQgkh0zjn+xghZCchxMvsxwYcKyeEvE0I8RFC3gCQO+h7byWE1BJCOggh9w46togQsosQ0kMIaSGE/FTy58smhKwnhLQRQrrY58UDjnsIIX9gP18XIeRFQkgagFcBTGaunF5CyOSBLhdCyGuEkC8NutZ+Qsin2OeUEDKNEHIHgLUAvsHO8zIh5D8JIc8P+t5fEkJ+LvhZRO/fc4SQ0+y930wImcler2LvmWXA2BsIIftG8t4qxhFKqfpQH9IfAE4BWDXE698HsA1APoA8AO8C+AE7dgWA0wBmAnAB+BMACmCawTU8ALoA3ArAAuCz7Oscdvw9AD8FYAewDIAPwJPs2AwAvex1OxsXTdwz+95b2efpAJYMuO4BADcb3FMOgBvY/bsBPAfgxQHH/w7gWQDZAKwAlrPXLwHQMOhc9w24388BeGfAsRkAugHY2df97xOAJwD8cMDYQgB+AFnsawuAVgALBL9Dw/ePHb+N/Yx2AD8HsG/AsWoAVw74+gUAXxO9t+rj7PwY9xtQH+fWB2cBOAHgqgFfXw7gFPv8cQAPDjg2TbAA3Apgx6DX3gPwzwBK2YSeNuDY0wMm1P8C8MyAY2kAwgMWgM0Avgcgd4Tvw4UAutjnhQDiALKHGCdaANxsEi9jX98P4PEBYw0XAPbaqwBuZ5+vBlAtuG/u+zfE+Cx2D5ns628CeIp97gHQB6Awme+t+hi7D+UCUiSLyQBqB3xdy15LHKsfcKz/c0JI6QD3SK/BuRLnK2LHuiil/kHHBt5H//nZuI4Bx78AoBLAB8y1tFrmhyOEuAghv2WupR5ok10WIcQMoARAJ6W0S+ZcA6GU+qDtHm5iL90E4KlhnOL/ANzCPr8F2u6KB/f9I4SYCSE/IoScYD/nKXYo4SZ6EsA1hJB0ADcC2EIpbWbHzui9VYwfagFQJIsmAGUDvi5lrwFAM4DiAcdKEp9QSuuolumSTilNNzhX4nyN7FzZzL8+8FiC5oHnJ4S4oLlvEtc7Rin9LDRX1f8A+MugcxnxNQDnAVhMKc2A5joBAAJtwfEQQrKG+D4Zud0/A/gsIWQpACeAtwzGDXWuFwHMIYTMgrYDEC0eovfvZgDXAlgFIBPAFPY6AQBKaSO03dj10HZq/QvOCN5bxTihFgDFmWAlhDgGfFigTWLfIYTkEUJyobliErnl6wB8nhByAZuQ/0tw/lcAVBJCbiaEWAgha6D5xtdTSmsB7ALwPUKIjRByEYBrBnzvXwCsJoRcRAixQYtN9P+dE0JuIYTkUUrj0HztABCT+JndAAIAugkhHgD/nTjAnoBfBfAICxZbCSGJBaIFQA4hJFPw85axe32W3dtQtADQ1QRQSoPsZ34amtusjvdDSLx/bgAhaLsmF4AHhjjNHwF8A8BsaDEAACN6bxXjhFoAFGfCK9Amw8THfQB+CG1iOQDgIIA97DVQSl8F8DC0J9vj0J4gAW2i+QiU0g5oT7NfgzYRfQPAakppOxtyM4DFADqhTcR/HPC9hwH8G7QJsRla8HhgHv4VAA4zd9MvANzEJlEQQg4TQtYa/Mw/h/Z03g4t2P3aoOO3AogA+ABaIPY/2P18AG1xrCGEdBNCJg/6PlBKQwD+Cu2p+2mD6wPA/wMwg53nxQGv/x+0yVjk/klg+P6xz2uh7baqof2sg3kB2oL1wiBXkuF7qzg7IZSqhjCKsYUQcgGAQ9AyXaLjfT/nOoSQUmgLzyRKac8YXfMEgDsppRvG4nqK0UHtABRjAiHkeuZyyIbmH35ZTf4jhxBiAvBVaJlPYzX53wAtHvHmWFxPMXpYxEMUiqRwJ7Q0xhiAtwH867jezQSABVhboLlsrhh0rHfIb9Jy+LeM4JqboMVjbuXEKhTnCMoFpFAoFCmKcgEpFApFinJOuYByc3PplClTxvs2FAqF4pxi9+7d7ZTSvMGvn1MLwJQpU7Br167xvg2FQqE4pyCEDK6sB6BcQAqFQpGyqAVAoVAoUhS1ACgUCkWKohYAhUKhSFHUAqBQKBQpyrguAISQxwkhrYSQQ+N5HwqFQpGKjPcO4AkMKmFXKBQKxdgwrgsApXQzNEnas4Km7gCe3VmHpu7AeN+KQqFQjDrjvQMQQgi5gxCyixCyq62tbVSvteVYG17e34wtx0b3OgqFQnE2cNZXAlNKfwfgdwCwcOHCUVWuu3h6ns4qFArFROasXwDGkslZTqypKhUPVCgUignAWe8CSgbKt69QKBQfZbzTQP8MrT/seYSQBkLIF0bjOsq3r1AoFB9lXF1AlNLPjsV1KgvcmJTZicoC91hcTqFQKM4JUsIFdLTFh9PeEI62+LjjlKtIoVCkEikRBJbN7km4igCoYLBCoZjwpMQCIJvdo1xFCoUilUgJF5AsG6pbsPloOzZUt4z3rSgUCsWooxaAAVQ3e9HpD6G62Tvet6JQKBSjjloABlDqSYPTZkapJ407TgWLFQrFRCAlFoA3Dp/GDb95B28cPs0dNznLgQK3A5OzHNxxqq5AoVBMBFIiCPzo5hPYV9+NRzefwGUzJxmOu2ZuEbJcNmG2kNIMUigUE4GU2AHctWwqLizJwl3LpnLHHW70Yt2uehxu5McAZMfJoNxJCoVivEiJBWBmUSZuXFiCmUWZ3HH3vXQIu2u7cd9L/AZlA3cUI0W5kxQKxXiREgvAo5tO4Ifrq/HoJv6E3d4X0lkjZHcUe+u68PXn9mFvXZfhmIun5+GauYVCd5LaKSgUimSTEjGAvfWd6A3HsLee33ysKMOJmo4AijKcSbnuU9tr8QarKZhXmj3kGNkiNVWlrFAokk1K7ADuXlGJ+aVZuHtFJXec1WLSWSMefLUau2u78eCr1dxxaxeX4bIZBVi7uMxwjMwuAZDfKSgUCoUsKbEDkI0BNHcHddYYMsgOTUGGA1VTPCjIME4rldklAKpZjUKhSD4psQN4dNNx/GB9NR7ddJw7joDqrBFrFpbA47JizcIS7riX9zfisc01eHl/o+GYy2dMwrT8dFw+wzg9VaFQKEaDlFgAXj90Gr2hGF4/xC8Em5rv1lkjttV0wBeMYFtNB3dcTyCKnmAUPYGo4Zia9l54+yKoae/lnkuhUCiSTUosAF2BsM4acay1V2eN6AlGEY1rloeMo8jHFgkfZ5FQKBSK0SAlFgCr2aSzRkTjcZ01wmY2gTDLo6atFx29IdS0GS8oGU4rMp1WZDit3HPJBotl00VVWqlCoUiJBSDdbtZZI66aWQgTszwmZzlgNROhZtD+hm7EqGaNqCr3YG5JJqrKPdxzJYLFT22v5Y6TLSyTHacWCoVi4pISWUBdfRGdNeLNIy2IM8ujyRtEJE7R5OVnC608vwB/3lmHlecXGI7ZUH0abx9pQ4Hbzs0CunzGJJxs9wuDxbI6RapLmkKhSIkFoDLfjUPNPlQKgrtdgZjOGtHQ1Yc41SyP7ac6EY1r1oi9dd3o9Iext854lwAAnX1hOK0WdPbx4xiy6aKy45TwnUIxcUmJBaCLTZpdgslTltPsyf+0YAcQicVAmTWC0g8/eIxXu0pVf6BQTFxSIgbQy7J1egVZO7Lkue06a0Qi/ZOXBmq3mkCZ5XG0xYfT3hCOtviGd7MKhUJhQEosACU5Lp0dKbE41VkjpuengzBrxKHGHlBmeXhcNgQiUXhcNu64ZAdtVRBYoZi4pMQCkG636KwRmXaTzhrR4QvrrBGdfWFQZo0ozLTrrBGvV5/G8dZevF7NL2aTze6RTStVctUKxcQlJWIAB1njloOCBi6hGNVZI6KDrBFWk0lnh2J5ZT5afGEsr8znnquqzIOdJztRVcZPFwUF6jv9EKhZ4LEtJ/DmB63oC0fxyNqFhuNUEFihmLikxA4gw2nRWSPizKUTF7h2ZDnC/PVHOH77Eo8LTqsJJR6+e+q1Q81o6g7gtUPN3HFPbqtFfVcAT27j1wu4rBZQqlkeiSDw5Cy+RLZyFSkU5x7jugAQQq4ghBwhhBwnhHxrtK4jE4wFgHBcb0eKxUx0dih+/dYx1HYG8Ou3jnHPFYzGEaea5ZGTbgOlmuXRF47CRAj6wskJjMu4itQioVCcXYzbAkAIMQP4NYArAcwA8FlCyIzRuFaETZoRweQpJ/IsTzBCdXYovKw4zSsoUqvMd8NlMwtrGeo7+0CZ5XH7sqm4ek4hbhd0NZOdtLU0VTs3TVXFExSKs4vx3AEsAnCcUlpDKQ0DeAbAtaNxoTQW/E0TBIHTbSadNcI0yBrhsBKdHYoCJidRIJCVONrSA18ohqMt/GyhTKYplCnQFpLpVQDISVpr9ydOU5VZJBQKxdgxngtAEYD6AV83sNd0EELuIITsIoTsams7syfH7DSrzhoRiMR11oj4IGtEhsOqs0ORm2aHiWiWxx6WrbNHkLVTkZcOh8WEijzj1FMAeGrbKfz49SN4atsp7jgCAkIIiGBfJNOxTLaWQbmKFIqxYTwXgKFmlI/4Siilv6OULqSULszLO7NMlDuXTUOZx4k7l03jjotSvR0p7SxNtJ2TLhqKxBCnmuVRlO3UWSMauwMIx+JoFEyee2q70dEbxp5avgTF6rmT8cWLy7F67mTuOJlgsewOQLmKFIqxYTwXgAYAA1tqFQNoGo0LvX2kFQ1dAbx9pHU0Tm+I2ay3Q3GwwauzRqSxbJ00QdZOhz+EONUsD4fNDLOZwGHjK6Qmkx0nO7Cvrhs7TvIb6aj+xwrF2DCeC8BOANMJIeWEEBuAmwC8NBoXer26BTGqWR4JNQaBKoM0hOjtUKQ5LDprRHWzT2eNOM2e/E8LdgA3LyrFnOJM3LyIr/OTzMIymQ5pgHzqqUKhGBnjtgBQSqMAvgTgdQDvA1hHKT08GtdK1HUJ6rsQi+vtSAlF9XYonCzg7BQEnrNY/CJLEMegg6wR++q7UNfRh331/JhCXbsfe2o7Udfu546T6VeQ7CwrhUIxMsa1DoBS+gqltJJSOpVSev9oXUd24pEN7spmAcnQ1hPWWSOiLIU1KkhlnZzp1FkjZGWo/7yzHn2ROP68s547bu3iMlw2owBrF5cZjnE7LchwWOAWFOTJooLFCsXISIlK4AuLMwAAS8qNG64MB9kFRWahsJj11ghZV5FsymtHbwgxqlkeZaxCuUxQqSyTVrqoPAcXlmZhUXkO91yyE7sKFisUIyMlFoBjbVpRVF1XAEdOj1xOOTbIGiGzo0goVIuUqhMqoCI1UEK0hYkXdwCAOlYoVicoGJPVUZKpF5BNA5WtPVDBYoViZEz4BaC+sw+9zAnf2B3E5T/fjE/+aiue3l4HX5BffWuEeZAdCbI++/dZ8Pd9QRA4HImDMssjwgIiEUFgxM12HG7BzkOmXkA2DVS29kAW5SpSKIZmwquBlnhcyHSY4Q3GYCJAnAIHGrw40HAQ319/GFfNLsRNVaWomiLvHpLdASQTQgBQ8ZN9R19IZ43Iz7ChyRtGfgZ/RxFgWkEBgWZQeW4aMpwWlOemGY7ZebIT++u9qMzv5PY/Xj13MjJd1qT1K1Z9jRWKoZnwCwAAOKzaApCbZsXDNy/AMzvq8Nqh0whG4vjrnkb8dU8jpiSpWcxwsUCTlRb9IuwWE6KROOwW/qYtzKrYwoJqNofFAiDMrDHpDiuC/gjSOdXMAPDC3kZUN/Xghb2NuGzm0I3rKSgopaDC/Y4cslLVStJaoRialFgAEhBCsKQiB0sqctATjOClfU14dmc9DjZ6capD7wt/o7oFl56XB4t5dL1kTrsJvlAcTkETGj9z6fgFrh1vIKKzRpxf6EZtZx/OL+S7YzIcFrT7I8gQuIA6/SEEInF0cgrQrplbhCyXTTgRr9/fhOd2N8DbF8Edy/lidTKovsYKxdCkxAJgNhGdBTR9nluWlOGWJWV4v7kHz+6oxxPvneo/fvsfdyE33YZPLyjGjQtLhNo6Z4ovFNfZkZJYHwTrBFq8WsVwi5fvKuroDeusEQcaunV2KA43erFuVz08Lhu3yEt2p/D0tlqs292A+o4+fP2K87ljFQrFR5nwQWAAcDC3icPAfXJBYQbuu3am7jUCoL03jEffrsGKn7yNT//mXTy/uwGBsLznP42pgKZx1ECTTaKeTFBXhoZuTTa6oZufBSSLi0lKuDjSEj969X3sru3Gj159n3uua+YW4fZlFbhm7ke0AXXUtPfCF4ygpr13+Dc8BCpYrEg1UmIH0MZy3dsEOe8D2fLNS/Hcrnqs29WAZm8Qu2q7sKu2C//9knyxch/rA9DH6QeQbGSb2gTZQhYULGi97HivYNykTCfa/T5M4hSgJbKxenml0ZB32dy+bCpcdgu3+AzQZCqe2l6LtYvLuMFnFSxWpBopsQNIdHgcTqfH4mwXvnLZedj6zRX4422LcNXsSbCayUcmr8e3nkSXf2j3iGyK59mMrIxGorMYr8OYhbngLCb+jki2Yb1sTwMZmQpA1RUoUo+U2AHYzCb4EYftDAK6ZhPBsso8LKvMQ6c/jBf3NuL766v7j39/fTUefPV9XDZjEm6qKsFF03JhEkxwuvNDSycdO01ODW8oprNGuG0m+MJxuAU+pfqOgM4OhSfNhkZvCJ40fuqpbMN62WDx5TMm4WS7H5fPGDo7KYEKFitSjZTYAfhZma1fUG4rKvDypNlw20XlutfSbGZEYhSvHGzG5x7fgY//z5v42RtH0dDVJ1UwZjLp7VghK2fRy3xJvQKfkkwvBdmAcm6aA1aTCblp/Cf7xq4ATnsDaOzi++w7+8JwWi3o7ONfVyYGoOIEiolESiwAYaq3RsgWeA3U+Nn5nVX4yWfmYmGZ5ltu9gbxi43HcPH/vAXCBrodxm+zbNaOLLJCdRkOs86OlHSWxprOSWdt8YV01ghCABBx0VtbbxDROEVbb5A7LpmNaJT+kGIikRIuIKeVIBChcCYpG2egxo/LZsENC4pxw4Ji1LT1Yt2uBvxldz3ae8NICHd2B+O476XDWFNVggsKM5JyD0bIxh2C0ZjOjvR8oERvh6DM40RNRwBlHr5S6dHTPviDMRwV6DbJBoE3VJ/G20faUOC2c4PA2kLRyV0oVFGZYiKREjuAGIv+xoYTBT4DKvLS8a0rz8e2e1bisc/pfddPvHsKV/5iC6755VY8ua0WPWeoQyRiXkmmzhoh06sAABKiogJxUSlV0x4Wb+gRxB1O9wQQZzYZnGzX9KBOtvNTXmXE6lSzGsVEIiV2AIl2u4K2u0nDYjbhshkFutfKclyo7ejDwUYvDjZ68cP11bhqdmHSr53wh4v84rLYzCaEouIAepQtrlHOIutirdZcgpZrTd1BnTUikd0DgPtkv7wyD4ebvFheqSQjFIqBpMQOIMNu1tnxYNPXL8G6O5fihvlFcFpNCEbj+OveD+WOzSagpYc/4cnQwgKsLYJAqyx+VqHsF1Qq25mmkJ2jLeRjQXifIBgvKtxLcPmMSZiWny7M7lm3qw51nQGs21XHHadQpBopsQDkpNt0djwghGBRuQc/ufFC7Lh3FR64fjbmFH/oponFgY89+CZue2InXj98GpFk9aUcIVaz3hoh42bLc9t11oiPTc3VWSNOtvvRE4jipKBdZU1bLyizPGT7EIwHKvtIMRqkxAIgm344VrgdVty8uBQvfekiXQP6GKV484NW3Pmn3Vj64EY8+Mr7OCGYtEYb2SBwIBLV2aHwsYCDTxB42HqiXWeN8AbC6O4Lwxvg/14r8tJBmOUh04dAdiJO9oStso8Uo0FKLABW5kqwClwK40FCoM5uBn5983xcPD23X4fot5trsPInb+OG37w7bveXSBISJAshnUWJ0znR4mYmPNcsEKDLS7frrBG9wRiCkRh6g/ybu3FhKUo9Tty4kF/kJdPTQHYiXr+/Cb/fchLr9zdxx8miqpQVo8HZNyOOArF4XGdHipno7UiIssqpWAy4ek4h/vSFxdj6rRX4yqpKTM7SCqF21+olEfbWdYHSsRGYSKyZorWz3R/R2ZFgZW+sVfAG76nrhC8Uw566Tu64J7edQm1nAE9uO8Ud9/SOOhxo8OLpHcaxAtmagmT3PlDZR4rRICWygPxMyMw/DCVPHuk2M7yhGNI5ypeyDNU3uCjLiS+vmo5/XzEN757owDM767D+QHP/8esfeReVBem4cWEJPjW/WCet4DATBGMUjmSsTkB/LUN0DEMSCbeJyH0iu6OoYemfNYI0UG8gjEiMcl1Ksl3NZHsfKBTjSUrsANJsFp0dKbIKmQn/Pi/rkSfJYDIRXDQ9F7+6eT5cFv24oy29+OHf38eiBzbgX57cjbePtiEWpwgx1baQQL0tURMnqo2TaWyfdBIlwIJS4FJWUFYqKCxbNCUbhFkeDosZZqJZI5L9ZK9QjCcpsQDINjaXRVYhUyaCKhtk7Yt+OO6lL30caxeXIt1uQTRG8eqh0/inx3fgov95U/p8GS6rzp5N5LptOmtEfWefzhrhtJphIprlMS0/HU6bGdPyjYPFi8pzcGFpFhaV53DPJRsrUNk9ivEkJRaAS88vgNthxqXnF4gHJ5FEGwBeO4DiLLvOyjCnOAv3Xz8bO+9dhZ+tmYvF5R4Amg7RQP62rxFBg+q3Duar7xD47BNeriR4u6TpY5XCfYKKYcmNAnac6kSMapZHcbYLhZlOFGcb94fecbID++q6seNkB/dcskFbld2jGE9SIgZw3bwi+MNRXDeP32Eq2TitQCCiWSNG0jPAaTPj+nnFuH5eMU61+7FuVz0e2XSi//iXn9mHTKcV188rwo0LSzBj8vB1iMwmAsSorp3mmZJmJfBHqLBDmrcvorNGyC5ihZkOtPsjKMzkq4vKZAE1dQfR4gsKq5RlUdXHivEkJXYAMhovw0FG5hnQJv+BdigaukM6e6ZMyU3DN644H4m1hkBrvOINRPDEu6dw1cNbsPqXW/CnbbXChvEDkdVRklEhlXWdhdmAsGCgieqtETnpdhBmeTy66Tj21Hbj0U3HDcfUdfgRDMdQ18EvPpN9sm/pCWLnqU5hFbhyFSlGg3FZAAghnyGEHCaExAkhxh0/koTHZUMgEoXHlZxKYBnhM1lk5ZtlSUztFMB796zEPVeejwr2RHuosQffffEQFt2/Qfp8iTi3KIGqIMOms0MRZCmvQV7TAABOG9FZIzysstsjqPA+3OgFZZZHTbvWJ5mXLeRJs8NECDxp/MVE9m/usc01+PuBZjy2uYY7TrmKFKPBeO0ADgH4FIDNY3Gx16tP43hrL16vPs0dJ9skJaGLdgYNxj5CTrpVZ5NJntuOO5dPxcavLcdzdy3FZxYUw2k1IzQop/PXbx0fsQ5Rc09YZ0dCiAVNQoJeyolWnEYtORPIZjKls94I6ZweCYcauxGKURxq7Oae6/GtNdhb243Ht/In9rx0OywmIix6k60/UCiGw7gsAJTS9ymlR8bqemsXl+GyGQVC3XiZtE0AWD69QGeNkFlQWnsjOjtSrIMsoOkQVU3x4KHPzMXO76zCjz41W/c9D71+BEsf3Ijb/rADrx3S6xA5LXo7EmQXWFlXkUyQHQA6WYygUxAr6GUidb0csboG5oJpELhiuvoioMzymJztQEGGA5Oz+fGJZLsxFQrgHIgBEELuIITsIoTsams7s+2vbPNw2e5cslo1MliI3o6U6CA7mHS7BTct0ksiZLusiFPgzSNtuOvJ3VjygKZDdLy1FwF2ooCgb4DM4uliwV+XIAgsGxiXHZdoZyxoa4z8DLvODsWcokydNeJrnzgP88uy8LVPnMcdd83cIty+rALXzOUnKKgdgGI0GLUFgBCygRByaIiPa4dzHkrp7yilCymlC/PyzixTItkqjz6WnugTpCm6mA/bxfFlJ9OdBJxZVtH2b6/CI2vnY3llHkwE6PBrOkSrfvq29DlkFk8/e1T3ix7Zk4yL6RO5BF1t0u0WEPD1jBJ9hUX9hTv9YbT7QugUuKdkg8Cy6acqWKwYDqOWBkopXTVa5x4ujV1BtPQE0djF/ycrybajriuEkmy+P9ZuBkIxzfIIsSbEIU4zYtkg62his5hw1exCXDW7EE3dAfxldwPW7apHw6CmMt96/gBurCrBvJIsEFHy/VmEbPvL3lAMlFkjatr8OmvEwxuPodEbxMMbj2HNImMRuse2nMCbH7SiLxzFI2uN8yFklEqBD4PFALCmii9+p1CkRB3A8VYf+sIxHG/l+0+dNguAELPGTMt343CzD9Py+dtx2SbzycRmAsJxsbvDiMlZTty9cjq+dOk0vFfTgbW/395/7Jmd9XhmZz0q89NxY1UJrp9X1J9ameW0oDsQRVYSggVmaO9ZsmrPEoJ7UUH2UUOnX2eHPJekNlJ+hh2N3iDXnQQA5TnpyHB0ozyHL1W9eu5kZLqswnoBVVegGA7jlQZ6PSGkAcBSAH8nhLw+mtezW0wwEc3yaO0J6azxuKDOGiHjjhlJIdhQOJkT3imKZAswmQg+Pi23/wnBBGAGa2h/tFXTIVry4Eb8y5O78daRVvQE5Lp9yZDshdPCAiwWQaClj7mm+jguKgeLXzgEcYzEw4booePCkiyUeJy4sCSLO04WWZeSDMqdNPdD/nEAACAASURBVPEZryygFyilxZRSO6W0gFJ6+Whe7+5Vlbh2XhHuXlXJHedm6X9uThogAKQzTaF0gbaQbFZRMulhrRt7BC0cZSEDYhSvfPlirP/3i3DrkjK4HRZEmA7R5/+wsz/FUqiPJEGaVW9HfD5JMUCZxVh2N5FItR2ccjsY2RRl2ToA2boCmcld1R5MfM76LKBkIJsF5GOTpk8wedrYTsKWhAYzyc4CSvaOIpERmrCzijLxg+tmYee9q/DzNRdiacVHRdFufmwbV4dIhEwF9XBwMSEjVxIEjSZl2nXWCDuL6tsF0X3Zvsay2kLluS6k2y0ozzXWMwLkJneVeTTxSYkYgGxgrCI3DXvquvsrZ41YdX4B6jtPYZVAXM5kAhBn1gDZ3HhZZAPUshgVUTmsZlw3rwjXzStCbYcfyx/a1H/s3RMdePdEBzIcFlw3rwhrqkowczI/bXIgmU4LugJRZCaj+ABAhsMKIMisMRYCRCl/MZbtQSArZyHb1zjh2qkscHObwqyaMQktvhBWSSwoA+1QDKw94PU+UJy7pMQOQPZJ5q7lUzG/LAt3LZ/KHbfxg1YEInFs/KB1xPcmW8wkq8pJ43o7FpTl6BfMK2ZOgsVE0BOM4o/v1eLqh7di9cNbpM/nD0d1dqS8f9qns0YkvDo8747MGABwMelpl0CCuicQgTcQQY9gu/PwxmP4274mPLzxGHdcomHNzpN85VOZWIFqQznxSYkFQLaKcn99N+o7A9hfzy/zj1OtHUhc0JYx0YEyGZ0oZXvzJjJOOZmno86jty7Ae/esxLevOh9T85gOUVNP/3GzCXj3eDviBgJzNKa3I0V2l5Uo1xBIEEkRZ064uMAZt7u2E62+EHbX8ifsYCSKaIwiGOEvirINa57aXos3qlvw1PZa7jjFxCYlXECyqXE17b3wBSOoae/ljvOHojprRGJ94K0T2U4zugIxZDv5T4oWlt55Fva1B6A9ScTx4RNFntuOO5ZNxe0XV2BPXRee3VmPdbsaAGjxhJt/vx2lHhc+s6AYn15YjMLMD90aLrvWctMl8GMNvqYRaayFZ5pg+2S2ECBCNWsAgRZfEa0RXtao3itoWL+H9XveM6jv82ASXcp43coA+VaUCVkUnjyKqimY+KTEApBoqC3i9mVT4bJbhJpBsoFWGU2bPhYo7RMETMNxvTXCqs1hwlaPycYoVkAIwYIyDxaUefoXgAR1nX34yRtH8bMNR7GsMg83VZVgxfkF/TETXuxkOPSxKrs+QbVdjP2iYpxfWE6aBe3+KHLSkvOvk+O2o8kbQo6bH1Q+yBRNDwoUTWX/1ueVZgv9+slW0VWcfZylz5Pjg2y20CTWWGSSoMFIOqvGSudUZSU2EYLNBHJYU+CENcLBck4dY5l7ega88ZVluP3icnjSbIhTYNORNtz15B4seXAjEs/XeW7++yur8ikbZ5FZZDv8UZ0dKbK7yUQAWxTIlkUmDfTp7XU4UN+Np7fXJeWairOPs3uWGGNk856XVuQi323H0opc7rg89lSXJ3i6k8HLmgInrBF9TIynT6Rol2SG29dgeoEb9149A9vuWYlHb5mPS8/TdIg6/WF0saKyo61+PLuzDr2i1XEMkd39lbAWnyWCVp89zEXUI3AVgVC9NUC2eOupbafw49eP4KltpwzHBKNRxKhmFROTlHABySIbK1i7pAylOS7huCtnFeIP757ClbMKR3xvJqaPYBJkAVlMQCwmjhXI+s9lSbOZ4AvHkTZMDQqbxYQrZhXiilmFaPYG8PzuBvz4H0f7j3/z+YP43svVWD2nEGuqSjC/NHtcdYjc7Od0C37OetbhrV7Q6S2DxTsyBPGOArcDNe0BFAh2RU9tO4V1uxpQ1+HHf15xgeG4k+196A1FcZLT/KYo0wWrqRtFmfyaAsW5i9oBDCDhP+XlWQ+HzcfbEYjEsfn4yGWj4zG9NcLGgoQ2QbDQTPR2pMhWvvIozHTiSyum616zW0zoC8ewblcDbvjNe7jsp5vxu80n0N47shaaQyHT6tPC3jBLkt64XhaX6BXEJw6wBjQHBI1oqpt96A5EUN3Mz3ibW5wJp9WMucXG9RknO3oRjlGc7OAnRSjOXdQCcAbIuorOK3DDbjHhPE79QTYrdsoWFD3J+rtlpapl/eKy/Y9lMp7OhB33rsIPrp2JWayh/fG2XjzwygdY8sDG/jHJ+iM2m/V2KPzsffUL3t9cpmORK9CzkK3vkO2SFgxHEYlRBAU1FH/d24DOvgj+urfBcEybLwTKrGJiohaAM0C2QGZ6QTqKs52YXmCs9LjyggI4LCasvIBfVSzre5bNeZeVoJA9n+yCMlwynVbcunQK1t+t6RD909IyZDgsiA6oIYgD+PHrR1DXYezOkMHJdk1Ozu5JNhtL9vflslt11og57El9DueJHQD2NXTrrBGeNDvMBNzexkVZLpiYVUxMVAzgDJBNtZPJyXbZzLCYiVCnRnZCyWIyCiJZZrdDG+cWCNqJOoyNJbOKMjGrKBP3XHUBXj98Gl9+Zl//sV+9dRy/eus4llbkYE1VCa6YNQkOVoUrm7sfZduXaBK2MV2s/WSXoA1lBzveIRjX0N2ns0bYzCYEIjHYBBpE111YhKbuAK670LgT2W0XlSMcj+O2i8q552rqDmDLsTZcPD0vae5TxdggtQMghPxJ5jWFHpmYQntvGLE4RXsvv3OUrM8+i7kcsgSuh0SmTZeo12MSGW6mkBEOqxnXDpq4Eim579V04D+e3YdF92/Ad188hEOCvPmBhFn8IjyCOEYCWZedLK2+iM4aEWCxhIAgpvD20Ta0+kJ4+6ixG1NWp0jWJbq3rgtff24f9tbxi94UY4fs/+LMgV8QQswAFiT/dlKP8hwXCLM8Zkxy66wRp71BnTXCMsgaIbvwyPi8kz0pDuSdb67AHz5fhStnTYLVrOkQ/WlbLVb/cqv07kmmrWU2kwrPFkiGy75vsr8HWWRdVE6rCfE45faNkJWVkNXaUvITZx/cBYAQcg8hxAdgDiGkh334ALQC+NuY3OEE58X9TeiLxPHi/ibuuHml2bCaIKzeDDMnfFjgjJd17Uimn6OrL6KzY43ZRHDpefn4zS0LsO2elfjO1Rdgev5HYy93/3kv3jHQIUpUT/OqqMPs+8IGOkYJZKrAAfnfg+zuSTZmc6jJi1CM4lCT8Q6pIjcdmS4rKnL53cpk+xVXlXmQ5bSiqszDHaca0Ywd3L8nSumDlFI3gIcopRnsw00pzaGU3jNG9zihWVqRA7uFDKmrP5D1B5sRiWuWRwbz/WckSUpZdoKSnfDGgpx0O754cQX+8ZVl+Ou/fkx37KX9TVj7++1Y/tBbeHjjMd0kIxPIDrPtQXiMC+1kJ/ZE7EcUA0q4dXjunRf2NqC6qQcvcDKFAKAnEEVPMNrfFc6It4+2MrcTX0VXNaIZO7h/JYSQ8ymlHwB4jhAyf/BxSumeUbuzFGHtkjKA8EW5AGD17EL8eWcdVs/mF5VV5KZjT3238KnNatJcHWejYoSZaAuJyH0i+hkIIZg/aMe0oCwbu2u7UN8VwE8TOkTT87CmqgQuC9AX5XciM5kBRMUFeckmnRWMpQsKxnpYS84eQWtOq9mEcCwOKydYHAjHEIzEhfGEFm8Q3r4wWgRux4rcdOyqFf9tqr7GY4foMfGrAO4A8JMhjlEAK5J+RymGbNONhu4AQIhmOcSZx1YkQ+xJs6HFF4Yn7ewT+pLdTcj47Afz/L98DMdbfXhuVwP+srsBHf4w3j7apguG8hJyYjG9HSsS0uMiCfJslxXt/giyXfwkgJw0K/zhEHI4q93BJiZAx3ETAcCRlh5E4hRHWnq441bOKMBpXxArZ/BTnmWz7BQjh7sAUErvYPbSsbmd1EP2aScQjiIWowgICnwOs3/Ww4J/2p5QRGfPRZg6hrBIbTDT8t2456oL8PXLz8PG91uxblc9Nh1pxUC3/nW/fgdrqkpwzdzJSLd/+G8i2xAm2fhYVNcniO6WeFxo93tR4uEnFXQxTakujrZUIpVUlFK6es5kNHbXYPWcydxxL+5pxOuHTiPNauE+7Ki00rFDylFMCPnUEC97ARyklI68LVYKI/u0M7/MgxPtfZgvCKBNynCgriuISQJF09JsF460+FGazZ8o0mwm+M9A42ckyObtxwbZ4WI1m3DFrEm4YtYknPYGmRKpxr76buyr78b3X67G1XMKcVNVCRaUZUvf23hxoN6rs0aE2RYmzNnKdPWFddaIY229iFGKY218yYg9dZ3wh2PYU8dvfpPMPgSyi0mqLjqykcIvAFgK4C329SUAtgGoJIR8n1KqagJGmVUzCtDiC2KVYPtclOVEQ1cQRYI/4sXluWjqDmJxOV/RVBaZiVF28kx2Y3sZJmU6+nspmAF8cl4RXjnYjEAkhr/s1txFFXlp43Jvw0F2UUwIfPKEPrOcVgQiYWQ5+e4kl8WMvlAULoH+1N0rK/Ho5hO4axm/5Woy+xDILiap2vxG9rEuDuACSukNlNIbAMwAEAKwGMA3R+vmFB8i2+t1QZkHuW47Fgh2CkXZDhRkOFCUzd8pLJriAWGWh8zEmG436exYIau3k8j+iQH42ZoLNR2i62ZhdpGmQ1TTps+Y2fh+C6Kxsc0GShYyNRm9TO+oV6B79NyeekTimuXR6Q+j3RdCp5+/o3h6Rx0ONHjx9A5+HwKZdFFZ2ZZU7X8suwOYQiltGfB1K4BKSmknIeTcdSKfQ8gW5cgG2vbVduN4mx/7avmaMaWeNKTbu1DqSeOOk+lEFmD+64CoSkkS2R2F2aSlFZlNw3PcZDqtuHVJGW5dUobqph6s21WPJ9491X/8C/+3CwUZdtwwvxg3LizBlFz+e3Q2YWUtRnlZYHEa11njcXIB6offPIbG7iAefvMY1iziPWVrf+uifdbL+xvx3K4GdPeFcefyaUOOkXWxpmrgWXYB2EIIWQ/gOfb1pwFsJoSkAeDPIIqkINvrNVGUMz0/nRtoe7W6RWeNONTYjd5QDIcEMsQyJIKsghoqaWQXgAh7tI+MQKluxuQM3PfJmboFAABaekJ4ZNMJPLLpBBaXe7CmqiQp/R/OFNk+DzIVwyb2DpsE77DFZEYIMVgEubFzizPR4g1yJagB4OZFZegJRnHzIn5qtGz9gcIY2b34vwH4A4ALAcwD8H8A/o1S6lcZQmODbK8CAgJCCIjgn7aMuX7KBC6ghu4AKLM8LKwDjYXTiWZSpl1njUjEm0Vx50RgWhSgli1my2Q59pmcXPuBFblbvnEp7l4xDYVMh2j7yU58dd1+VN2/QXCl0SOZchsBll8bEOTZWlnBhlVQuOGyWmA2Ebis/OfOzr4wnFYLOgXBZ5niOKU/xEdqAaDafmwrgDcBbACwmb2mOMtYPXcyvnhxOVbP5afkfWf1TCwoy8J3Vs/kjku4uEWu7jTmYE/jONotzAVjEbhiZK8pWx4rK6Mg20shQYnHha9+4jxs/eYKPPH5Klw1W9MhGtzC8g/vnESXwO99NpKfYddZI2KsYCMmKNzYW9+FUIxibz1/Mk5mEPixzSfw9wPNeGzziRGfayIiqwZ6I4Ad0Fw/NwLYTgj59JlelBDyECHkA0LIAULIC4SQrDM9l+LMkFV6vP3icnhcVtx+MV8SOINlimRwMkZku2lJp3dKpuTIPhXLrCdDnctsIrjkvHw8svZDHaKBfO/laix6YAO+9PQebD02tA6RCFkpiFyXRWdHQmIzJ2ovameBBLugrDzLaQNhlsfjW2uwt7Ybj2+t4Y6TiYtV5KbD7RDrGcnqD000nSJZF9C9AKoopf9EKf0cgEUAvjuC674BYBaldA6AowCUrlCSkNVRkQ0qZ6XZMGNyJrIEFcNXzpwEl9WEK2dOMhzTzjpLtSepw5SfObD9SQoqj7SuAPhQh2ggaTYzIjGK9Qeaccv/246L//ct/HzDUTQOYxJxWPXWCH8kprMjQXonJkl+hh0mIt5RtPpCiDPLwxeMIRiJwRc0/llvXlKGr32iEjcv4ccTZP9vJppOkewCYBpU8NUxjO/9CJTSf1BKE/vkbQCKz/RcCj2y6WyySo/d/jCqm7zoFrgwXtjXiL5IHC/sazQcI9s3WLYNZcKTNMzknhEhm1I6kJ3fWYWHPj0HVVO0oHxjdwA/33AMF/3Pm9LnCET01gjZ1pEyNHpDOmt8bzGdNeJwkxcxKq5Sryxww0wglJdu6OpDMBpHQ9fIOsEB8v83stLX5wqyk/hrhJDXCSH/TAj5ZwB/B/BKku7hNgCvGh0khNxBCNlFCNnV1jYxVt3RRDZY/Hr1aRxv7cXr1ae54x5/5yQ6+yJ4/J2T3HEyVaOJ+UH0cJqoJxLUFUn18E02CV00gT6aDpfNgs8sLMFzd30MG7+2HHcur0Buuu0j/ZO//3I1jpzmN3MXkVBtEKg3JBk5X5yPCdT5BEJ1rT0hxKlmeRRnO2G3mFCcbfy3vn5/E36/5STWC+TWZf9vZKWvzxVkg8D/CeB3AOYAmAvgd5RSbgEYIWQDIeTQEB/XDhhzL7TkjKc41/4dpXQhpXRhXl5qFWmMJmsXl+GyGQVCFdLiLK1hTbGgL+yS8hwQZo2QzdqRnWQTsdaQIL1HtumKTD+AkTI1Lx33XHkB3rtnJX57q76n0uPvnMTlP9+Ma3+1FU9vr4MvOPwSm8SiMpYpGja2AtsEK3EiOD44SD6Yhu4+lnnGf7LvC8cQjVH0cf5QGrr60OwNJGWXAMhn2Z0rSEeKKKXPA3h+GONX8Y4TQv4JwGoAK1VG0dgzrzRb2FwGAL5zzQw8tb1WuFAsnZqLhu4glk41lpYwEaKzRjDFZaELSKb4DAAsFk3ywCL4a5fpByCbZy+qUbCaTbh8ULxkSo4Lpzr6sL/Bi/0NB/GD9dW4anYh1lSVCK42vgQjUZ01Qla9VbYC+UiLD6FoHEdajHdNhxt74A/FcLiRr1QqS1W5B0dbfagq51fGnyuI+gH4MPS+jkDLDs04k4sSQq6AJiGxnFKanKVZMSrILhSr505GpsvK9aGWelw42OxDqUCpUjZvX2bCBgCzSZu2NTsyZDOKzkQz6K2vX4LtJzuxbmc9Xjmk6RA9v6cBz+/hN2QZyHiolcruxGQxsXfXJHiXV88uRGNXgNsjo603CMosj711Xf0POry/d1n59nMFUUcw94BOYAM/3Gc6+TN+BcAN4A1CyD5CyKMjOJfiLEDGhzqvzAO33Yx5Ap0i2SCwLBa247AIdh6yqZajBSEESypy8FOmQ/TD62ZhzhBVs7f/cRc2VA+tQ+Rk2yHnaPqxBiG7KGY7zTprRB5Tss0TKNr+o/o0uvoi+AcnjiWbevzYlhNYf6AJj23h1wtMNM2gZPWiHhaU0qGFOxTnLDJyum6HGQ6rGW5BQ/UMpwVdgaiwraXdDIRimuUhW+BlJtqTM2+ukO2kZgUQYfZMyHBYccuSMtyypAzvN/fgyl9s6T/2RnUL3qhuQb7bjhsWaDpE5QkdIskggIyMhuz7K4u2A4sJd2Ih5iMKCXxFR1t6QZk1oqM3rLNG5KY5YDWZkJvGX3QmmmbQWdgQUHEuIpcfrQXQkiUInZC8Fklfyz6hynQiS6TNitJnLewJ3JKEJ/ELCvWb7Yun54IQLU/+N5tO4NIfb8KNv30Pz+9u6M+wEmTaSr3DifVSsjBaSHdfRGeN6GNNj/oEzY9kFjuH1ayzvGtGYlR8zSQz3oVl47IDUEw8ZDqb9QajCISj6BWkAUbYDBwRSAvIdLUCAIcFCEY1y8PCgsoWzpydKJwTFdAlMx9/MH/6wmI0dPXhuV0NeG53PZq6g9hxshM7BkiFUwpQStmC+1Fkg9nJRDY+4WeFXX5OgRegtb/0hUPc9pf5bjtafGHku/nFZzIBZUA+ViDL+v1NeG53A7x9Edyx3LhPwmg1rFE7AEVSkIkB1HX5EYzEUdfFl5/41LxiWE2a5ZGfYdNZIxLrjWDdQWGWXWeH4mirX2eN8LBeux5eh/kRUJztwlcuq8SWb6zAH29bhKvnFOrE2KIUuOIXW/D41qF1iJIpGpdsiElvjfCyX6iX84u1Mh0Lq0DP4u4V0zG/LAt3r5jOHffwhmP4295GPLzhGP/mJOkJROANRNAjqPB7eX8jHttcg5f3GxdanglqB6BICjJPKKXZaXBYu1CazdfNr+/uA0CYNcbjssMEPzwu/tOdrGx0Irc7GTnesXhcZ0cLs4lgWWUellXmocsfxoIfvNE/qR857cP311fjwVffxydmTsKahSW4aFouTCZyxv2UR0Km3QRvKI5MQUMgWQkKmbRih8UMM9Esj8tmTsJlHBmTfgj7SyLJ2dm5nRZkOCxwC+Jdo1V/oBYARVKQaalXlO3EpEwnijiVmwAQDMcRi1MEBRo/DqsJIMxykAnuAkCiHCUZZSmJeV80/8suTjLB5+wBek0EwGcXl+Jv+xrhD8Xw9wPN+PuBZkzOcuAzC0r6F4lkTGM21lxGJN/tDcV11giLiSAao0LVWJm+xjaLCXGqWR6yrh3ZXgWyyPb5GK36A+UCUiQFmfQ4WanqyVlOWM1E6Os82NiDONUsj0TWkSj76DSTHjjNkSCQTRX1ssipVxBBla0XkK3wHTixP3D9bOy69zL85DNzsYhNHE3dQfxi47H+68UBhKIji/Imu08yZU/XVPCUHWHBhAgnqPB+cw8oszx+sP4w/rK7ET9Yf5g77rVDp3GowYvXDvElVGT7EMhKUAysP0gmagFQJAXZP2QZOv1hxCkV9o5NNGJJWCOcNovOGiIxu49XU/gz7abmtJlxw4JirLtzKd76+iX4l0umIm9QQHTx/Rtx30uHhZOkETYWNbfxoufDIPFAz3mwByDX1czFVPtcAvW+uo4+nTXitcPNCMUoXjvczB332OYa1oeAL2kty2jVH6gFQDFmyErplua44LCZUZrDrxheVpmLvHQbllUay08AQLrdorNGzJ6cqbMjQbYJjSzJCNqW56bhm1ecj/e+tUL3encggifePYUrf7EF1/xyK57aXoueYegQFbidOjtSZLOFEqEEXkihoSugs0YsKvfAQtC/UzIiyGoTgoIaBafVhHicwilwT453GqhaABRjhuxTzOQsBwrcDkzO4j/ZX1iSjdIcFy4s4afjBVlyfFAgQ9rcE9DZkeBhDVk8SWjMkmwsZlO/v95KgG9ecT6msMX2YKMX975wCIuG0dYyygId0VEOeA8m4dfn+fdnscV8lmBRz023w2EzIzedn1BwwSS3zhrR5A0iEqdo8vIlKMa7D4FaABRjhqybaFF5Di4szcIijrIooKln7q3rFkpVyzrQZfLPZTtuyaQojieJB9goBf7lkql46+uXYN2dS3HD/GI4reaPPOE+suk4WnuGnswSqqVnol46FLJSIDINgU529OqsEVuPt6M3FMPW4+3ccbJFauFIDJRqlodsfwHlAlKkDLKa651+TTe+08/XjW9jsYQ2QUwhoVDAUyqQ7biVqMQVVeQmMpNEGUoJ93qS3OwfiWUQQrCo3IOf3DgXO+5diQeun60b/7+vHcHSH72JL/7fTvzj8GlEBuRodgViOjtSktGZLUG+26GzRtR39umsEVaWTmoVpJWGY1qpYFhQzPjinka8fug0XtzDz+9PZoxtIGoBUJx1yOY83/bxCpR5nLjt4xXccVOZbMNUgXyDzEQmW+Gb7OyeEo9TZ0cTt8OKmxfrU3mzXVbE4hQb3m/FHX/ajaUPbsSDr76Pmjb+k/VoIhMXqWMTep1gYnewldUhWGG7WcOjbk7jIwBoZq6fZoEL6LXDzfCFYsKg8mhx9jkoFSmPjLQ0AIAAJZ40YU7m8so81LT7sbySf74CtxUtvggK3MbVu26HGd5gTJhSKotscLeuI6CzY822b6/EhupWPLurHluOtaG9N4zfvl2D376dnCyX0UImVRQAfGGqs0bIuoA62G6zQ7DrTATbhxN0TyZqB6A465Dd7sr6RV/cr/UrflFQRl+R54aJaNYIK+u1aBX0XJR9opQlmW6RM8FuMePqOYX4422LsPWbK/CVVZVDivDd89cD2FvXNWQxnWwNRSJZS5C0JXW+ZL9vCQnuoaS4ByK7A7y0Mh9motnxQC0AinMW2YViSXku7GaCJeX8dNH5pdnITbdjPqcS1M+e/PyCJ8BMp1VnJxJFWU58edV0bPnGpfjTFxbpjv15Rz2uf+RdfOJnm/H7LTW6Wg7ZSVG2wUwyazJk24YGonprRCL7UyQbPrckC5lOK+aWZIlucVRQLiDFhOdrl5+HRRUe4U5h1YwCtPiCWDWjwHCMrM/exqK6NkF0V16nSG7cWGIyEVw8Pa9fWZQAOG+SGx+c9uFYay9++Pf38aPXPsDlMybhMwv5wn7jjWwXOoeZIBijcAh+rwtKs7HtVBcWCBRD//heLTr7Ivjje7W485Kxb5OidgCKCU9LTxA7T3WixSCNMcHG6hZsPtqOjdUthmOWVuSAMMtDRlYCGL/K4mQyUH7i1S9fjJe+9HGsXVyKdLsF0RjF3w8245//sDPp1012sZ0MQZbVExRk9+ys7dJZI+YUZ8JMMGTnt7FA7QAUE56nttfiDTap84S+atp74QtGUNNunNnitJlhIprlIatoKavKaWLjkjXZjdaOghCCOcVZmFOche9cPQOvHmrGszvrsX1ArwIAWPv7bVhTVYpPzCgQNmsxYjwkrc1EaxgkSts9U+mOsUYtAIoJz9rFZTprxO3LpsJlt3DH7W/wIkY1yyPPbUOLL4w8N79XgWyQ0mYlCEQobEnq9SuzoIzUPeW0mfGp+cX41PxinGr345Ifb+o/9s7xDrxzvAOZTiuun1eEGxeWYMbk4bUZt0Bz2fAmMdmfIT/ditbeCPLT+TEbm1nz/wvWf7isBP4IhUvw+9p5qhMxqtnxQC0AignPvNJsqe5NBRkOVE3xoIDTeeNKzwAAGvxJREFUjPzuFdPxyKbj+FeBvzbMKsDCgkow2QnKbjEjEInCLihAkkVm4ZGuZZAYNyU3rX+3Y4IWb3nzg1Z4mQ7RE++ewqyiDJ2UeDIahxZn2VHfHUIxp8kPoOkhDbRGhGN6a0SQ1YkEBfUiOek2tPsjyEnnPygkuxNZArUAKBQMmZ4GnjQbct12eNL4/7D5GXZ0BaLIz+BPPDJtKAGAMl8CFfgU3HYzfKEY3IJO7uPRECYxZ8YB/O5zC9HmC+Gvexrw7M561LT7caixB4caD/WPp+C3tZRZAJq9IZ01QnZil+kbDcjv7AIs1SkgSHn6rxcP4mCTD0eae/Dy3csEZ5VHBYEVCoZMXcGjm09gX303Ht18gnsuT5oNhFnuOPbk5xE8AcqmlcZpXGeNSMypnGZao06e2447l0/Fxq8tx3N3LcWnF2g6RANZ/tAm/Pqt40MG8BP9Ynh9Y2SVRWVJdi1DU3dIZ42oZTLVtQK56uGiFgCFgiFTV3DXsqm4sCQLdy0zbuANAJMzXbCZCSZn8iWtE8VUQxVVDaSN6R21CXSPIuzRNCJ4RHUyOVCnqI3XGEAIQdUUD378mbnY+Z1VumN1nX146PUjWPrgRtz2hx147dCHOkTJntxlkHWLedLsOmuEk+3UnIId279dOg0elxX/dmlyU0WVC0ihGAYzizJx48ISzCzip+119YURZ5ZHTZtfZ42IslkuKpjtEs29RE2+Sj1pONzsQ6mH359ZBtkAqgyDezZ84aJy/HVPA7r6InjzSBvePNKGnDQbblhwdtcVtDKXU6vA9RRkHeOCgs5xiypycKytF4sE6cfDZfyXf4XiHEJWl7042wWHxYTibP4OYOX5+TAxy+Pj03J11ghZddFEqisv5VWW1t6IziaT766ege3fXoVH1s7H8so8EKLp6/xuQKctMwH8orLhMUY2BhAZZI146LUP8Nc9jXjotQ9GdmODUDsAhWIYJOIDoqriomwHCjIcKMrmyxBvPdGOOLM89tZ36awRsi4Kh8WEQCQGh6BZugwyDeuHQ6KyOHE6m8WEq2YX4qrZhWjqDuAvuxuwbld9f5evGAWq7t+AT86djBurSjCvJMswcHyucqrDjzjVbDJRC4BCMQwScQIR18wtQpbLJlwouplWTrdANTLA+g8EBH0IKvLScLTVj4o8kWtHNpwpQZIrsniL2OQsJ+5eOR1funQaKr79Sv/rfeEYntlZj2d21mNafjpuqirB9fOKknNDZwEBlp4UEKUpDZNxcQERQn5ACDlACNlHCPkHIWTyeNyHQjHerDi/QGeNmMRqEyZxahQA+ebmsrLGMsi6MRI1UaJaNpldjGlQ6s/3PjkTFxRqKq7HmQ7R4gc2Cu5o+GQ5zTo7VvSyznK9Se4wN14xgIcopXMopRcCWA/gv8bpPhSKUUE2VtDgDYAwyyPEWjSGBM3IZbVqEvILPBkGp0VvR8poZu3808em4NUvL8P6f78Ity4pg9thQXRQzcRP/nHEsOOXrK5QN2sW1J2k7mfjzbi4gCilPQO+TMO5rYWlUHwE2VjBeQVuHGvpxXmCnrCBSFRnjajM11xAlfl8F1CcTY5xTmGZbNGTLGMhfDerKBOzijJx79UX4LVDp/Efz+7rP/bLN4/jl28ex9KKHNy0qASXz5zUvwCOh67QcEh3WNAViCLdkdwpe9xiAISQ+wF8DoAXwKWccXcAuAMASkvFvleF4mxANlZQkGGH22FBgaBiON1hRU8ohHQHP9XSbjbrrBEuuxW+cAgu+8hTN62smlnk2hlLSWuH1Yzr5hXhq8/u65/U8912tPpCeK+mA+/VdCDDYcG1FxZhTVVJ0q+f7Eprl82MrkAULpEI0TAZNRcQIWQDIeTQEB/XAgCl9F5KaQmApwB8yeg8lNLfUUoXUkoX5uUJWgQqFOcYsv2PP1ahNbX5WAU/DfRUZ6/OGpGoO+LVH8mGiU0mvTViPKSvBz7Rv/utFXj8nxfi8pkFsJgIeoJR/GlbLVb/cmvSr5vwrJ2h0OlHmJThBGE2mYzaDoBSuko8CgDwNIC/A/jv0boXheJsJcNpRabTigyBxMOkTAcyXTZMyuQHgc0s/dEsSINsZ1lH7ZzsI4fVjFAsJpRrdtrMCAViQons8cZiNmHF+QVYcX4B2nwhvLC3Ac/srP9IEd6Xn9mLNVUlWFKe85FgsyzJdp/NLMrE0VafsABxuIyLC4gQMp1Seox9+UkAya1uUCjOEVbPnYxMl1UYK5BdKC4szcbbR9txoUAxssTjwpEWP0o8xoVq3mBMZ43oYQHRnnMoMJrntuOOZVNx+8UV2FPXhRt+817/sb/ta8Lf9jWhJNuJNVUl+PSCEuHCO5hErF4Qs5cmiUm7OsYrBvAjQsh50HZotQDuGqf7UCjGFdlYQVW5B0dbfagq93DHhaNxmIhYhnp2URZqOwKYXWTci1ZW4iE7zYoOfwTZafxxToumpZ+srKJkQAjBgjL9ezqvNAt767pR3xXAj/9xFD994yiWTc/DTYuSHyuQpb6rD8FoHPVdyRWDG68soBvG47oKxbnKjpMd2FfXjen56Vw9+GA0hhjVLI/Grj6Eo3E0ciaUTn9EZ42QySgCgDS7FYFoBGlJCDwDmtJmKMpX3BxcVSzDC//6cRxr8WHdrno8v6cRnf4wNh1tw6aj/JTe0aSrL4xIjAq1pYaL0gJSKM4BZIPFmU4rLCaxbDQI0aSgObECWbeD22HWWSM62ELSIVhQZEkoXvOUr2UDz4OL1KYXuHHv1TOw7Z6VePQWTYdocDjg+l+/g2d31o1Ih0hWNvog60B3UNCJbricRZsxhUJhhGys4O6VlchJtwvbX2axBSKLt1BI5m16+6I6a0SG3QxvKIYMgfSxNBIrlIn18BXFchONuwY38LJZTLhiViGumFWIZm8ASx98s//Y3vpu7K3vxn0vVWP1nELctKgE80uzh6VDlCjEFhVkm81ANKrZZKIWAIXiHEA2ViDb/rKmXRMXq2k3FhezmbUgpii5J9NlhTcUQ6aLv+soyHLA2+JHQRY/oFqYaUOzN4zCTH6THJkuXsnMxinM1KdgXnvhZLx68DQCkRie292A53Y3YGpeGm6S+D0lkN2huO02hKJhuO3892S4KBeQQpGCLC7PQbrdjMXlxvryBWzCK8jk556vXVwGj8sq3HV09oZ11nAcS03tFAjkjQcDNx2/uGkedt67Ct+/diZmFGoN7U+0+XH/K+/rvicaS0Yq0OhUUagFQKFIQYqyHLCaTSjiPI33sAbpPYJG6ZuOtqK7L4JNR1u543yhiM4akXCpn2US/wA+Og1nuqz43NIpeOXLF/frEGUMkmv42I/exP++9gFqRyDl7GNvhi/Jb4paABSKFOTxd06hsy+Cx985ZTjGz7pU+QXdqjr9EVCIs4UK3A6dPZuQbaTDY1ZRJn5w3SzsuFdfA9vqC+GRTSew/KFNuPG37+GFvQ0IMlnvhHdN5NqPMR9WLFmVZQy1ACgUKUiiUQ2vYU2u26azRiye4oHFpNlzlUQGqyCTVYrBldP/vmJav4z3jpOd+Mqz+1H1ww34zgsHpTuHzS3O1NlkoRYAhSIF+e7qmfj0giJ8d/VMwzFpNisIszxePtCESFyzPBK+8OT4xOWQlXmeztRTpwtUVM+Er33iPLzzrRX4w+ercMXMSbCYCHyhKJ7cXqcb5+0z3kF1svz/ziTXAagsIIUiBZHJFrp+XhF+u7lG2FkrHo/rrBGne8I6a0Rumg3t/jBy00ae8WKzAMGoZnmMduDZbCK49Lx8XHpePtp7Q3hxbyOe2VGP420fivZV3b8Bn5hZgM8uKsXSCr0OUV1nQGeThdoBKBSKITnW1osYpTjWxlcW/fj0PFiIZnnIau57mPSERyBBIYdcOVs7i1+0J6lIjUduuh1fvLgCb3x1me6uwrE41h9oxtrfb8eyh97CLzYcRVO3NuEnWjcnoYWzDrUDUCgUQ5JI6xSld5bnpiE7zYbyXL77RLZ5vJ+1PfQL2h/aTEA4rlkjIqz9WGQ02pCNEEIIXFYCf4TCZQHuu3YO/ryzDnvrutHQFcDPNhzDLzYew8en5cJhtSAUiyLLldw6ALUAKBSKIZEtKmvpCcEXjKKlJ8Qd57KZ4Q3GhE1NEq0cB7d0HAylejsUskHW8aKPlR4HosCNVSW4saoEx1t9WLerAX/Z3YBOfxhbjrX3j7dZkqsHqlxACoViRBxp8SEUjeNIi487LlEslbBGxNjEHxMsABY2GVqSPCnykNVHkk3vHKq8a1q+G9++6gJs//ZKPHrLAlxy3oeuta4ku6jUAqBQKEbE3SumY35ZFu5eMZ07rpNluXRysl0AoDcY0Vkj0lhkN00U4U0isvW4ydh5WM0mXDFrEp74/CLMKtR6RlcKekcPF+UCUigUI+KymZNw2cxJwnFLKnLQ5A1gSYWx/AQgp/EDfOgOSbZbZCwxM7E6UQFaZ5/mXjvdo7KAFArFOciMwgx4XDahCyjPbdVZIxL9cZPdJ3cskRWr8waiOpss1AKgUCjGhCe3nUJ9VwBPbjvFHZeoExPVi80qyoLbbsYsTlezTCY9nSmQoJb12Se6o4m6pCUbE5OYNg1DalrqvEk9m0KhUBgwv9SDNJsZ80v5khGzizJBmOXhdljgtFngdhh7si0scd4iSKBPhBFE4YTW3ojOjhVutoC5k9VLgaFiAAqFYky485KpmFGUIWxqE4rGQYhmefQGowiEo+jl1AvIKpomu11lsmlnEhDtSgpCoVCci8g2tQlGY4hL9DXecqwVvlAMW44Zy1AnPCZiz4lcfo9kk7SkYzOZEEYcNlNynTbKBaRQKM4q7BYzTESzPOq7Ajo7FJX5bp01IiHPnLBGjE5bFjH5brvOJgu1ACgUirOK6fnpcNnMmJ6fzh1XmGnX2aGo7+zTWSPCLA0nLEjHsQyyRsiqkMriZzmxflFu7DBRC4BCoTircDuscNkscDv4/vi8dCcIs0ZE4zGdNUK29kC2wCuP9VDIE/RSkK0s7mWdwHpVRzCFQjGRyXBakOGwIMPJf862WQkIs0Zkuew6a4TdordGyPYNCIajOjtSZCW3h4taABQKxVlFRW46Ml1WVOTyXUALSj3IdduxgJNWOrc4C2aiWR6y7SqjbAKO/v/27j5Grqu84/j3t+u1vV5Hu/Fr7Pg1rh0S7NqJ3ECb4DgEVxSFBhAQKFJp8kdqKdSAFLWUCkqpkKKWvkkIUHDTUBSoLRITEioM5aVx/yj4JdnYcSAbEuO3dez43Xjjl92nf9yzdNf23JldT7g7e38faXVm7ty5e47Ozj5zz7n3PFX+EZ9Nr5+tsl/Ncwq1z2YPia8CMrMRZePOA7x48BQbdx7IXWLij948l9mTJ+ReVtpzvpcmiZ4qVxSdSWP/Z6rMAbQ0Nw0qK/7ec4PLSmq9qii7ASx8I5iZjW4fetNcVl0/vWoeglpMnjCOJmVlnluumUxTKvOcSwHiXLVAocFlJbXOAbSmPMOtLfW9EazQACDpfkkhaUqR9TCzkeOGOVfy+fctq5qLYFPXIZ7o7GZT16GK++w73sO5vmDf8fxF1LbsPkpfKvNMT5dhTq9yOWZ/zoNquQ9qzZJW6/GGqrAAIGk2sArYXW1fM7MLvWXhVN65dEbuEND4liaaJcZXSUN2x5IZTGhp4o4lM3L36zp4alBZyexJEwaVldS6BtHVHa2Dynopcg7gn4A/Bx4vsA5m1qBqubN4ze2LmDzxl1WHk57ec4wz5/t4es+x3P0mtDQD51JZWVdKjtNVJUlOrX524MSgsl4KCQCS/hDYFxGdqjKpIele4F6AOXOq30ZuZtav1rSWh391lt7Iyjwn03X4J6tcj19LukqAMWOg93xW5mkdO4ajPb201jn5zes2BCTpvyTtuMTPncBfAZ+u5TgR8WBELI+I5VOn5i8iZWY2HPfcPJ+5k1q55+b5uftdM2UiSmWeWZNaB5WVnDs/uKxkbhpKmltlSGmoXrcAEBFvi4jFF/4ALwHzgU5Ju4BZwDZJ1VMKmZm9HgSzJ7VVvRxn9coF3Di3g9UrF+Tud1V7K03KyjzXpVSP/WUlO/YdH1TWy298EjgitkfEtIiYFxHzgL3AjRFx4DddFzMzyHLtXtU+rmrO3c49x9hzpIfOKnMFEUFfZGWeJg0uKzmbsuOcrZYlZ4h8H4CZld4Lr5zkwPEzvFBl0nZn9wmOnj7Lzu78ydgd+08MKivZffj0oLKSeVPaBpX1UvidwOkswMysMP2XklZLVjN+TBN9fcH4KhnGZrSP5+TBXzGjPX9piTFjmuBMb9WMZYtndrDr1dMsnpm/pMVQFR4AzMyKVmuymj1HTxOpzNN9/LVBZSXnU9az81Wyn/Wc66WpSfRUyVcwVB4CMjOr0ZrbF3HDnA7W3L4od79bfmsKzcrKPKfO9A4qK7l10VSmXTGOWxfV90pInwGYmdVo1Ruvyl2grt/Ka6exs/sEK6+dlrtf27gmTpzpo21c/nfxjc8dYP/x19j43AHuuql+90P5DMDMrM7Wb93D3mM9rN+6J3e/+25byKQJLdx328Lc/XrOnqe3N+ipU36Bfg4AZmZ1tnrFApbN7mD1ivz7BTraxnL9zHY62vIzh83smEBLs5jZUd8bwTwEZGZWZ7UOFdV69ZEngc3MSmrprHZaW5pZOqu9rsd1ADAzK8iTnftZu+llnuzcn7vfui27OXL6HOu21Hf1fAcAM7OCBEFEENWyAocGl3XiOQAzs4K8c+nVdEwYW3UO4C/fcR1ffuoXVSeVh0rVFisaSZYvXx5btmwpuhpmZg1F0taIWH7hdg8BmZmNcPuP9bBu8272H8vPbTxUDgBmZiPcpq5DPNHZzaauQ3U9rucAzMxGuFrvFxgqBwAzsxGu1tVKh8pDQGZmJeUAYGZWUg4AZmYl5QBgZlZSDgBmZiXlAGBmVlIOAGZmJdVQawFJOgT8cphvnwK8WsfqFMFtGBlGQxtgdLTDbajN3Ii46C6yhgoAl0PSlksthtRI3IaRYTS0AUZHO9yGy+MhIDOzknIAMDMrqTIFgAeLrkAduA0jw2hoA4yOdrgNl6E0cwBmZjZYmc4AzMxsAAcAM7OSKkUAkPR2ST+X9KKkTxRdn+GQtEvSdknPSGqIxMiSHpJ0UNKOAdsmSfq+pK5UXllkHaup0IbPSNqX+uIZSe8oso7VSJot6UeSnpf0nKSPpu0N0xc5bWiYvpA0XtJPJXWmNvxN2l5YP4z6OQBJzcALwCpgL7AZ+GBE7Cy0YkMkaRewPCIa5qYXSSuAU8C/R8TitO3vgCMR8UAKxldGxF8UWc88FdrwGeBURHy+yLrVStIMYEZEbJN0BbAVeBfwJzRIX+S04f00SF9IEtAWEacktQD/A3wUeA8F9UMZzgBuAl6MiJci4izwH8CdBdepFCLiKeDIBZvvBL6aHn+V7EM8YlVoQ0OJiO6I2JYenwSeB66mgfoipw0NIzKn0tOW9BMU2A9lCABXA3sGPN9Lg/3hJAF8T9JWSfcWXZnLMD0iuiH7UAPTCq7PcH1E0rNpiGjEDp1cSNI84AbgJzRoX1zQBmigvpDULOkZ4CDw/YgotB/KEAB0iW2NOO51c0TcCPwBcF8amrBifAlYACwDuoF/KLY6tZE0EXgU+FhEnCi6PsNxiTY0VF9ERG9ELANmATdJWlxkfcoQAPYCswc8nwXsL6guwxYR+1N5ENhANrTViF5J47n947oHC67PkEXEK+mD3Ad8hQboizTm/CjwSEQ8ljY3VF9cqg2N2BcAEXEM+DHwdgrshzIEgM3AQknzJY0FPgB8u+A6DYmktjTxhaQ24PeBHfnvGrG+DXw4Pf4w8HiBdRmW/g9r8m5GeF+kycd/BZ6PiH8c8FLD9EWlNjRSX0iaKqkjPW4F3gb8jAL7YdRfBQSQLg37Z6AZeCgiPldwlYZE0jVk3/oBxgBfb4Q2SPoGsJJsudtXgL8GvgWsB+YAu4H3RcSInWSt0IaVZEMOAewC/rR/DHckknQLsAnYDvSlzZ8kG0NviL7IacMHaZC+kPTbZJO8zWRfvtdHxGclTaagfihFADAzs4uVYQjIzMwuwQHAzKykHADMzErKAcDMrKQcAMzMSsoBwMyspBwArHQkrZT0ewOer5b0x1Xes1bS9enxJ+tUj4clvbcexzIbDt8HYKVzucs5SzoVERPrUI+HgScj4puXeyyz4fAZgI0akr6VVkt9rn/FVGXJgLalJBw/SCtJrgY+nhKIvCUlFblf0nWSfjrgePMkPZse/1jSckkPAK3pvY9I+tv+5CRpv89JWlOhfpL0BUk7JX2HAas+Svq0pM2Sdkh6MO27QNK2AfsslLQ1PX4gHedZSSN+LXwbmcYUXQGzOronIo6kdVY2S3qcbIGwFRHxsqRJ6fUvM+AMQNLtABHxvKSxkq6JiJeAu8hu0f+1iPiEpI+kFR37lyZ+DPgXSU1ka01VWpDs3cC1wBJgOrATeCi99oWI+Gw65teAOyLiCUnHJS2LiGeAu4GHJU1Kx3pDRET/+jJmQ+UzABtN1kjqBP6XbAXYe4GnIuJlgBrXV1lPlmUKsgCwLm/niNgFHJZ0A9kifU9HxOEKu68AvpFWr9wP/HDAa7dJ+omk7cBbgTem7WuBu5VltrsL+DpwAngNWCvpPcDpGtpldhEHABsVJK0kW13xdyNiKfA00MnQcz+sA94vaRFZEqeuGt6zliy94t38/zf6Si6qj6TxwBeB90bEErKzlvHp5UfJckDcAWyNiMMRcZ7sLONRsuxR362hjmYXcQCw0aIdOBoRpyW9AXgzMA64VdJ8yJJvp31PAldc6iAR8QugF/gUlb/9n0tr0/fbQLau++8AG3Pq+BTwgZQVagZwW9re/8/+1ZTw5NdXBkXEa+mYXwL+LbVjItAeEf8JfIxsNUyzIfMcgI0W3wVWp0nbn5MNAx0iGwZ6LI3PHwRWAU8A35R0J/BnlzjWOuDvgfkVfteDwLOStkXEhyLirKQfAcciojenjhvIhne2Ay8A/w1ZchBJX0nbd5HlsBjoEbLE4d9Lz68AHk9nDgI+nvM7zSryZaBmlykFl21k67jXMmQ01OPfT/aN/1P1PraVm88AzC5DujnsSWDD6/TPfwNZztu31vvYZj4DMKszSUuAr12w+UxEvKmI+phV4gBgZlZSvgrIzKykHADMzErKAcDMrKQcAMzMSur/AMxxQDRXiBCJAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot regplot of `activity_days` log-odds\n",
    "sns.regplot(x='activity_days', y='logit', data=logit_data, scatter_kws={'s': 2, 'alpha': 0.5})\n",
    "plt.title('Log-odds: activity_days');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cp7ojoBldEYy"
   },
   "source": [
    "<img src='images/Execute.png' width='100' height='100' align=left>\n",
    "\n",
    "## **PACE: Execute**\n",
    "\n",
    "Consider the questions in your PACE Strategy Document to reflect on the Execute stage."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "F_l3bkxQdJ3a"
   },
   "source": [
    "### **Task 4a. Results and evaluation**\n",
    "\n",
    "If the logistic assumptions are met, the model results can be appropriately interpreted.\n",
    "\n",
    "Use the code block below to make predictions on the test data.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "id": "cSl5gbXfBPBN"
   },
   "outputs": [],
   "source": [
    "# Generate predictions on X_test\n",
    "y_preds = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "dbN_LHUb9uhp"
   },
   "source": [
    "Now, use the `score()` method on the model with `X_test` and `y_test` as its two arguments. The default score in scikit-learn is **accuracy**.  What is the accuracy of your model?\n",
    "\n",
    "*Consider:  Is accuracy the best metric to use to evaluate this model?*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 27,
     "status": "ok",
     "timestamp": 1685661318649,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "fU3v-XO49qm8",
    "outputId": "935ee835-3272-4f5c-f842-6e0979ee273f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8237762237762237"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Score the model (accuracy) on the test data\n",
    "model.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "mwRmSDS3eyeH"
   },
   "source": [
    "### **Task 4b. Show results with a confusion matrix**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "qUKLVt50-zFE"
   },
   "source": [
    "Use the `confusion_matrix` function to obtain a confusion matrix. Use `y_test` and `y_preds` as arguments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "id": "IBFU_dicBjwQ"
   },
   "outputs": [],
   "source": [
    "cm = confusion_matrix(y_test, y_preds)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EKS_DsNcGotN"
   },
   "source": [
    "Next, use the `ConfusionMatrixDisplay()` function to display the confusion matrix from the above cell, passing the confusion matrix you just created as its argument."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 449
    },
    "executionInfo": {
     "elapsed": 26,
     "status": "ok",
     "timestamp": 1685661318650,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "BT0hOH_ZFsnx",
    "outputId": "b032d12d-97c3-4f64-a6c6-4ed062f97fdd"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm, \n",
    "                              display_labels=['retained', 'churned'],\n",
    "                              )\n",
    "disp.plot();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "iYR3wnJvteDg"
   },
   "source": [
    "You can use the confusion matrix to compute precision and recall manually. You can also use scikit-learn's [`classification_report()`](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.classification_report.html) function to generate a table from `y_test` and `y_preds`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 26,
     "status": "ok",
     "timestamp": 1685661318651,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "bP3cDYQRgGwO",
    "outputId": "eaaeeb6e-aa06-417b-e353-7bd3cdd3471d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5178571428571429"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate precision manually\n",
    "precision = cm[1,1] / (cm[0, 1] + cm[1, 1])\n",
    "precision"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 21,
     "status": "ok",
     "timestamp": 1685661318651,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "hAz7NVP3fxQD",
    "outputId": "74c07f46-b040-4755-d8e4-f537f3752e55"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0914826498422713"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate recall manually\n",
    "recall = cm[1,1] / (cm[1, 0] + cm[1, 1])\n",
    "recall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 18,
     "status": "ok",
     "timestamp": 1685661318652,
     "user": {
      "displayName": "Matteo Riotto",
      "userId": "15211004813978485634"
     },
     "user_tz": 420
    },
    "id": "-RBqS0tLpv6p",
    "outputId": "3425e075-692e-48bc-e85a-dd3d17ee9d3c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "    retained       0.83      0.98      0.90      2941\n",
      "     churned       0.52      0.09      0.16       634\n",
      "\n",
      "    accuracy                           0.82      3575\n",
      "   macro avg       0.68      0.54      0.53      3575\n",
      "weighted avg       0.78      0.82      0.77      3575\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Create a classification report\n",
    "target_labels = ['retained', 'churned']\n",
    "print(classification_report(y_test, y_preds, target_names=target_labels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "PfZ5fuh_gPTH"
   },
   "source": [
    "**Note:** The model has mediocre precision and very low recall, which means that it makes a lot of false negative predictions and fails to capture users who will churn."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "gQfAkdAivpvU"
   },
   "source": [
    "### **Task 4c. Conclusion**\n",
    "\n",
    "Now that you've built your regression model, the next step is to share your findings with the Waze leadership team. Consider the following questions as you prepare to write your executive summary. Think about key points you may want to share with the team, and what information is most relevant to the user churn project.\n",
    "\n",
    "**Questions:**\n",
    "\n",
    "\n",
    "1. What variable most influenced the model's prediction? How? Was this surprising?\n",
    "\n",
    "> _`activity_days` was by far the most important feature in the model. It had a negative correlation with user churn. This was not surprising, as this variable was very strongly correlated with `driving_days`, which was known from EDA to have a negative correlation with churn._\n",
    "\n",
    "2. Were there any variables that you expected to be stronger predictors than they were?\n",
    "\n",
    "> _Yes. In previous EDA, user churn rate increased as the values in `km_per_driving_day` increased. The correlation heatmap here in this notebook revealed this variable to have the strongest positive correlation with churn of any of the predictor variables by a relatively large margin. In the model, it was the second-least-important variable._\n",
    "\n",
    "3. Why might a variable you thought to be important not be important in the model?\n",
    "\n",
    "> _In a multiple logistic regression model, features can interact with each other and these interactions can result in seemingly counterintuitive relationships. This is both a strength and a weakness of predictive models, as capturing these interactions typically makes a model more predictive while at the same time making the model more difficult to explain._\n",
    "\n",
    "4. Would you recommend that Waze use this model? Why or why not?\n",
    "\n",
    "> _It depends. What would the model be used for? If it's used to drive consequential business decisions, then no. The model is not a strong enough predictor, as made clear by its poor recall score. However, if the model is only being used to guide further exploratory efforts, then it can have value._\n",
    "\n",
    "5. What could you do to improve this model?\n",
    "\n",
    "> _New features could be engineered to try to generate better predictive signal, as they often do if you have domain knowledge. In the case of this model, one of the engineered features (`professional_driver`) was the third-most-predictive predictor. It could also be helpful to scale the predictor variables, and/or to reconstruct the model with different combinations of predictor variables to reduce noise from unpredictive features._\n",
    "\n",
    "6. What additional features would you like to have to help improve the model?\n",
    "\n",
    "> _It would be helpful to have drive-level information for each user (such as drive times, geographic locations, etc.). It would probably also be helpful to have more granular data to know how users interact with the app. For example, how often do they report or confirm road hazard alerts? Finally, it could be helpful to know the monthly count of unique starting and ending locations each driver inputs._"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "provenance": [
    {
     "file_id": "16ynSVRiYFz40jV9BFaXgY62vWJWEVjFA",
     "timestamp": 1671052218770
    },
    {
     "file_id": "15PPfvGb4OuUkuQoTTpKKgSeY9o9a_XrL",
     "timestamp": 1669649625868
    },
    {
     "file_id": "1_uBujapIzHItho27E-iPg9wn3aHtsjLG",
     "timestamp": 1664565386285
    }
   ]
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}