LendingClub_DefaultModeling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "de42e925",
   "metadata": {},
   "source": [
    "##### High Level Outcomes (also shown below, placed here for convenience)\n",
    "* Compared to the performance of the loans in the testing dataset, this model would increase annualized returns by 38.5% (% change from 1.92% to 2.65%).  \n",
    "\n",
    "* If we were to extrapolate to the results on the test dataset to the overall loan portfolio, we could estimate the following: \n",
    "\n",
    "    * 11.2% default rate \n",
    "\n",
    "    * 275M in returns vs. 3.4B in capital  \n",
    "\n",
    "    * The model would reduce gross returns by $80M but require nearly 2.5B less capital to do so  \n",
    "\n",
    "* This model would outperform the proposed intervention of setting an interest cap of 13.5%, because it would deliver about 7M more in revenue, while also using about 17M less capital.  \n",
    "\n",
    "* The company would generate smaller returns in absolute terms if they used this model, 22% less in absolute terms, but their % returns would increase because they would generate those returns via deploying 45% less capital. Another way to look at is that the capital cost to generate returns from their loan portfolio would be nearly cut in half.  \n",
    "\n",
    "* This model has the highest annualized returns compared to the two interventions proposed in the EDA Notebook: \n",
    "\n",
    "    * No loans below Grade D, 2.2% annualized return \n",
    "\n",
    "    * Setting an interest rate cap of 13.5%, 2.4% annualized return \n",
    "\n",
    "* The value this model would have for the company would depend on a lot of factors: namely, are they going to optimize capital efficiency or gross returns? Also, the company could elect to make up the difference with growth and marketing. I.e., aggressively seek out additional customers who meet their new lending standards. \n",
    "\n",
    "* The model and the suggested interventions from the data analysis say the same thing: after a certain point capital efficiency drops off substantially and the company is investing a considerable sum of money for a small gain.  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84bfa650",
   "metadata": {},
   "source": [
    "##### Key Notes:\n",
    "* The key evaluation statistic will be the financial performance of the loans chosen by the model, rather than F1 or AUC. That said, F1 was the metric used for cross-fold validation and for determining the most valuable features.  \n",
    "* XG Boost was the \"winning model\", I have included an L1 Logistical Model from Sklearn just as a reference. The performance of both models was quite close, but XG Boost ran in 7 seconds on a GPU vs. 11 minutes for SK Learn on CPU, as the latter lacks GPU acceleration capability and XG Boost is just faster in general. An L2 regressor and a decision tree were also tried, but not included in this notebook due to the L1 giving better results.  \n",
    "* Used under sampling techniques to deal with the data imbalance (more paid off loans than defaults) rather than using techniques that would \"simulate\" extra data  \n",
    "* The hyperparameters were chosen by grid searches done separately from this analysis  \n",
    "* Predictions made based on the \"paid\" column, namely, if the customer paid off their loan the column has a one, if they defaulted, it has zero.  \n",
    "* Given this is an imbalanced dataset, the key metric targeted as the F1 score as it's a balance between the true positive rate and what % of loans that were positive did the model classify as good. E.g. If the model is right 85% of the time when it comes to classifying good loans, but only recognizes 70% of the loans as good, the true value of the model is somewhere between those two numbers. \n",
    "\n",
    "\n",
    "#### Key statistics on the loans in the original dataset\n",
    "* Default Rate: 19.18% \n",
    "* Gains from good loans: ~912m\n",
    "* Losses from bad loans: ~560m\n",
    "* Net Gains: ~353m \n",
    "* The  loans in the bottom two \"grades\" are unprofitable overall, with the grade above it barely profitable (i.e. grades E, F, G) \n",
    "\n",
    "#### Financial Targets for the machine learning model \n",
    "* The model is attempting to beat one or both of the potential interventions revealed in the EDA portion of this analysis, namely:\n",
    "    * Not originating loans to customers with loan grades below D, meaning: <16.5% default rate, >351M in gains, <5.09B in deployed capital, annualized return of 2.2% \n",
    "    * Setting an interest cap at 13.5%, <3.6B in deployed capital, >270M in gains, <12.3% default rate. 2.4% annualized return\n",
    "\n",
    "The general idea is that the model doesn't per se have value from being just accurate, instead the loan portfolio identified by the model has to outperform the typical \"human identified\" interventions to improve portfolio performance. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc3e37a2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt \n",
    "from eda_class import EDA\n",
    "import warnings \n",
    "warnings.filterwarnings('ignore')\n",
    "%matplotlib inline\n",
    "import xgboost as xgb\n",
    "from sklearn import linear_model, metrics\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.model_selection import GridSearchCV \n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "%matplotlib inline\n",
    "from xgboost import plot_importance\n",
    "from numpy import mean\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedStratifiedKFold\n",
    "from xgboost import XGBClassifier\n",
    "from numpy import sort\n",
    "from sklearn.metrics import f1_score\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from numpy import loadtxt\n",
    "from xgboost import plot_importance \n",
    "from matplotlib import pyplot\n",
    "warnings.simplefilter(action='ignore', category=FutureWarning)\n",
    "from sklearn.metrics import classification_report, confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "33c5e3b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# set size parameters for the plots \n",
    "\n",
    "plt.rcParams['figure.figsize'] = [12, 8]\n",
    "plt.rcParams['figure.dpi'] = 100 \n",
    "\n",
    "# set view parameters for the pandas data frames \n",
    "pd.options.display.max_rows = 999\n",
    "pd.options.display.max_columns = 999\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "33c77866",
   "metadata": {},
   "outputs": [],
   "source": [
    "# instantiate the EDA functions that will help us analyze the data \n",
    "\n",
    "eda_functions = EDA()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b7dce581",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.56 s, sys: 140 ms, total: 1.7 s\n",
      "Wall time: 1.7 s\n"
     ]
    },
    {
     "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>funded_amnt</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>installment</th>\n",
       "      <th>delinq_2yrs</th>\n",
       "      <th>fico_range_low</th>\n",
       "      <th>inq_last_6mths</th>\n",
       "      <th>mths_since_last_delinq</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>collections_12_mths_ex_med</th>\n",
       "      <th>acc_now_delinq</th>\n",
       "      <th>chargeoff_within_12_mths</th>\n",
       "      <th>acc_open_past_24mths</th>\n",
       "      <th>avg_cur_bal</th>\n",
       "      <th>delinq_amnt</th>\n",
       "      <th>mo_sin_old_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_tl</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>mths_since_recent_inq</th>\n",
       "      <th>mths_since_recent_revol_delinq</th>\n",
       "      <th>num_accts_ever_120_pd</th>\n",
       "      <th>num_il_tl</th>\n",
       "      <th>num_tl_120dpd_2m</th>\n",
       "      <th>pct_tl_nvr_dlq</th>\n",
       "      <th>tot_coll_amt</th>\n",
       "      <th>tot_cur_bal</th>\n",
       "      <th>total_bal_ex_mort</th>\n",
       "      <th>emp_length_months</th>\n",
       "      <th>dti_dec</th>\n",
       "      <th>length_of_credit_history</th>\n",
       "      <th>total_payments</th>\n",
       "      <th>post_loan_dti</th>\n",
       "      <th>net_gain</th>\n",
       "      <th>debt_consolidation</th>\n",
       "      <th>consumer_credit</th>\n",
       "      <th>other</th>\n",
       "      <th>countdown_zero_delinq</th>\n",
       "      <th>countdown_zero_revol_delinq</th>\n",
       "      <th>term_ 36 months</th>\n",
       "      <th>term_ 60 months</th>\n",
       "      <th>grade_A</th>\n",
       "      <th>grade_B</th>\n",
       "      <th>grade_C</th>\n",
       "      <th>grade_D</th>\n",
       "      <th>grade_E</th>\n",
       "      <th>grade_F</th>\n",
       "      <th>grade_G</th>\n",
       "      <th>home_ownership_ANY</th>\n",
       "      <th>home_ownership_MORTGAGE</th>\n",
       "      <th>home_ownership_OWN</th>\n",
       "      <th>home_ownership_RENT</th>\n",
       "      <th>verification_status_Not Verified</th>\n",
       "      <th>verification_status_Source Verified</th>\n",
       "      <th>verification_status_Verified</th>\n",
       "      <th>paid</th>\n",
       "      <th>application_type_Individual</th>\n",
       "      <th>application_type_Joint App</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20000.0</td>\n",
       "      <td>0.1485</td>\n",
       "      <td>691.84</td>\n",
       "      <td>0.0</td>\n",
       "      <td>690.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.845</td>\n",
       "      <td>12.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5356.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>102.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>37491.0</td>\n",
       "      <td>37491.0</td>\n",
       "      <td>72</td>\n",
       "      <td>0.1245</td>\n",
       "      <td>102</td>\n",
       "      <td>35.975100</td>\n",
       "      <td>0.199973</td>\n",
       "      <td>4889.01</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20000.0</td>\n",
       "      <td>0.1577</td>\n",
       "      <td>700.88</td>\n",
       "      <td>0.0</td>\n",
       "      <td>680.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.366</td>\n",
       "      <td>63.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2759.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>137.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>39.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>92.1</td>\n",
       "      <td>264.0</td>\n",
       "      <td>63456.0</td>\n",
       "      <td>63456.0</td>\n",
       "      <td>60</td>\n",
       "      <td>0.2221</td>\n",
       "      <td>137</td>\n",
       "      <td>31.076057</td>\n",
       "      <td>0.342251</td>\n",
       "      <td>1780.59</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-46.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>0.1797</td>\n",
       "      <td>253.78</td>\n",
       "      <td>0.0</td>\n",
       "      <td>685.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.780</td>\n",
       "      <td>28.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>20578.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>176.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>288087.0</td>\n",
       "      <td>71518.0</td>\n",
       "      <td>24</td>\n",
       "      <td>0.3570</td>\n",
       "      <td>176</td>\n",
       "      <td>21.901647</td>\n",
       "      <td>0.412370</td>\n",
       "      <td>-4441.80</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>20000.0</td>\n",
       "      <td>0.0849</td>\n",
       "      <td>631.26</td>\n",
       "      <td>1.0</td>\n",
       "      <td>705.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.057</td>\n",
       "      <td>15.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>17700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>93.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>141601.0</td>\n",
       "      <td>27937.0</td>\n",
       "      <td>120</td>\n",
       "      <td>0.1761</td>\n",
       "      <td>201</td>\n",
       "      <td>34.119870</td>\n",
       "      <td>0.265219</td>\n",
       "      <td>1538.51</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-81.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>0.0649</td>\n",
       "      <td>306.45</td>\n",
       "      <td>0.0</td>\n",
       "      <td>685.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.345</td>\n",
       "      <td>23.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1997.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>95.7</td>\n",
       "      <td>8341.0</td>\n",
       "      <td>27957.0</td>\n",
       "      <td>27957.0</td>\n",
       "      <td>72</td>\n",
       "      <td>0.1307</td>\n",
       "      <td>164</td>\n",
       "      <td>35.891570</td>\n",
       "      <td>0.173964</td>\n",
       "      <td>998.97</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   funded_amnt  int_rate  installment  delinq_2yrs  fico_range_low  \\\n",
       "0      20000.0    0.1485       691.84          0.0           690.0   \n",
       "1      20000.0    0.1577       700.88          0.0           680.0   \n",
       "2      10000.0    0.1797       253.78          0.0           685.0   \n",
       "3      20000.0    0.0849       631.26          1.0           705.0   \n",
       "4      10000.0    0.0649       306.45          0.0           685.0   \n",
       "\n",
       "   inq_last_6mths  mths_since_last_delinq  open_acc  pub_rec  \\\n",
       "0             0.0                     0.0       8.0      0.0   \n",
       "1             0.0                    38.0      24.0      0.0   \n",
       "2             0.0                     0.0      14.0      0.0   \n",
       "3             0.0                     3.0       8.0      0.0   \n",
       "4             1.0                     0.0      14.0      1.0   \n",
       "\n",
       "   pub_rec_bankruptcies  revol_util  total_acc  collections_12_mths_ex_med  \\\n",
       "0                   0.0       0.845       12.0                         0.0   \n",
       "1                   0.0       0.366       63.0                         0.0   \n",
       "2                   0.0       0.780       28.0                         0.0   \n",
       "3                   0.0       0.057       15.0                         0.0   \n",
       "4                   1.0       0.345       23.0                         0.0   \n",
       "\n",
       "   acc_now_delinq  chargeoff_within_12_mths  acc_open_past_24mths  \\\n",
       "0             0.0                       0.0                   3.0   \n",
       "1             0.0                       0.0                   9.0   \n",
       "2             0.0                       0.0                   4.0   \n",
       "3             0.0                       0.0                   4.0   \n",
       "4             0.0                       0.0                   7.0   \n",
       "\n",
       "   avg_cur_bal  delinq_amnt  mo_sin_old_rev_tl_op  mo_sin_rcnt_rev_tl_op  \\\n",
       "0       5356.0          0.0                 102.0                   16.0   \n",
       "1       2759.0          0.0                 137.0                    2.0   \n",
       "2      20578.0          0.0                 176.0                    7.0   \n",
       "3      17700.0          0.0                  55.0                   32.0   \n",
       "4       1997.0          0.0                 129.0                    1.0   \n",
       "\n",
       "   mo_sin_rcnt_tl  mort_acc  mths_since_recent_inq  \\\n",
       "0            10.0       0.0                    9.0   \n",
       "1             2.0       1.0                    9.0   \n",
       "2             6.0       6.0                    6.0   \n",
       "3            13.0       3.0                    8.0   \n",
       "4             1.0       1.0                    1.0   \n",
       "\n",
       "   mths_since_recent_revol_delinq  num_accts_ever_120_pd  num_il_tl  \\\n",
       "0                             0.0                    0.0        1.0   \n",
       "1                             0.0                    5.0       39.0   \n",
       "2                             0.0                    0.0        6.0   \n",
       "3                             0.0                    1.0        9.0   \n",
       "4                             0.0                    0.0        3.0   \n",
       "\n",
       "   num_tl_120dpd_2m  pct_tl_nvr_dlq  tot_coll_amt  tot_cur_bal  \\\n",
       "0               0.0           100.0           0.0      37491.0   \n",
       "1               0.0            92.1         264.0      63456.0   \n",
       "2               0.0           100.0           0.0     288087.0   \n",
       "3               0.0            93.3           0.0     141601.0   \n",
       "4               0.0            95.7        8341.0      27957.0   \n",
       "\n",
       "   total_bal_ex_mort  emp_length_months  dti_dec  length_of_credit_history  \\\n",
       "0            37491.0                 72   0.1245                       102   \n",
       "1            63456.0                 60   0.2221                       137   \n",
       "2            71518.0                 24   0.3570                       176   \n",
       "3            27937.0                120   0.1761                       201   \n",
       "4            27957.0                 72   0.1307                       164   \n",
       "\n",
       "   total_payments  post_loan_dti  net_gain  debt_consolidation  \\\n",
       "0       35.975100       0.199973   4889.01                   1   \n",
       "1       31.076057       0.342251   1780.59                   0   \n",
       "2       21.901647       0.412370  -4441.80                   1   \n",
       "3       34.119870       0.265219   1538.51                   0   \n",
       "4       35.891570       0.173964    998.97                   1   \n",
       "\n",
       "   consumer_credit  other  countdown_zero_delinq  countdown_zero_revol_delinq  \\\n",
       "0                0      0                    0.0                          0.0   \n",
       "1                1      0                  -46.0                          0.0   \n",
       "2                0      0                    0.0                          0.0   \n",
       "3                1      0                  -81.0                          0.0   \n",
       "4                0      0                    0.0                          0.0   \n",
       "\n",
       "   term_ 36 months  term_ 60 months  grade_A  grade_B  grade_C  grade_D  \\\n",
       "0                1                0        0        0        1        0   \n",
       "1                1                0        0        0        0        1   \n",
       "2                0                1        0        0        0        1   \n",
       "3                1                0        0        1        0        0   \n",
       "4                1                0        1        0        0        0   \n",
       "\n",
       "   grade_E  grade_F  grade_G  home_ownership_ANY  home_ownership_MORTGAGE  \\\n",
       "0        0        0        0                   0                        0   \n",
       "1        0        0        0                   0                        0   \n",
       "2        0        0        0                   0                        1   \n",
       "3        0        0        0                   0                        1   \n",
       "4        0        0        0                   0                        0   \n",
       "\n",
       "   home_ownership_OWN  home_ownership_RENT  verification_status_Not Verified  \\\n",
       "0                   0                    1                                 1   \n",
       "1                   0                    1                                 1   \n",
       "2                   0                    0                                 1   \n",
       "3                   0                    0                                 1   \n",
       "4                   0                    1                                 1   \n",
       "\n",
       "   verification_status_Source Verified  verification_status_Verified  paid  \\\n",
       "0                                    0                             0     1   \n",
       "1                                    0                             0     1   \n",
       "2                                    0                             0     0   \n",
       "3                                    0                             0     1   \n",
       "4                                    0                             0     1   \n",
       "\n",
       "   application_type_Individual  application_type_Joint App  \n",
       "0                            1                           0  \n",
       "1                            1                           0  \n",
       "2                            1                           0  \n",
       "3                            1                           0  \n",
       "4                            1                           0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# read in data \n",
    "\n",
    "data = pd.read_csv('data/updated_lc_ML_ready_data_april2022.csv')\n",
    "\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d04222d9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(386080, 61)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 386,080 rows, 60 columns \n",
    "\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b442ee14",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['post_loan_dti',\n",
       " 'application_type_Joint App',\n",
       " 'length_of_credit_history',\n",
       " 'tot_cur_bal',\n",
       " 'home_ownership_RENT',\n",
       " 'term_ 60 months',\n",
       " 'installment']"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# in order to avoid overfitting, identify and then remove highly correlated features, as these redundancies can \n",
    "# cause overfitting. We call the \"find_correlations\" function in the external EDA class to identify these features \n",
    "\n",
    "list_cor = eda_functions.find_correlations(data, 0.8)\n",
    "\n",
    "list_cor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e8b5651d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>funded_amnt</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>delinq_2yrs</th>\n",
       "      <th>fico_range_low</th>\n",
       "      <th>inq_last_6mths</th>\n",
       "      <th>mths_since_last_delinq</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>collections_12_mths_ex_med</th>\n",
       "      <th>acc_now_delinq</th>\n",
       "      <th>chargeoff_within_12_mths</th>\n",
       "      <th>acc_open_past_24mths</th>\n",
       "      <th>avg_cur_bal</th>\n",
       "      <th>delinq_amnt</th>\n",
       "      <th>mo_sin_old_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_tl</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>mths_since_recent_inq</th>\n",
       "      <th>mths_since_recent_revol_delinq</th>\n",
       "      <th>num_accts_ever_120_pd</th>\n",
       "      <th>num_il_tl</th>\n",
       "      <th>num_tl_120dpd_2m</th>\n",
       "      <th>pct_tl_nvr_dlq</th>\n",
       "      <th>tot_coll_amt</th>\n",
       "      <th>total_bal_ex_mort</th>\n",
       "      <th>emp_length_months</th>\n",
       "      <th>dti_dec</th>\n",
       "      <th>total_payments</th>\n",
       "      <th>net_gain</th>\n",
       "      <th>debt_consolidation</th>\n",
       "      <th>consumer_credit</th>\n",
       "      <th>other</th>\n",
       "      <th>countdown_zero_delinq</th>\n",
       "      <th>countdown_zero_revol_delinq</th>\n",
       "      <th>term_ 36 months</th>\n",
       "      <th>grade_A</th>\n",
       "      <th>grade_B</th>\n",
       "      <th>grade_C</th>\n",
       "      <th>grade_D</th>\n",
       "      <th>grade_E</th>\n",
       "      <th>grade_F</th>\n",
       "      <th>grade_G</th>\n",
       "      <th>home_ownership_ANY</th>\n",
       "      <th>home_ownership_MORTGAGE</th>\n",
       "      <th>home_ownership_OWN</th>\n",
       "      <th>verification_status_Not Verified</th>\n",
       "      <th>verification_status_Source Verified</th>\n",
       "      <th>verification_status_Verified</th>\n",
       "      <th>paid</th>\n",
       "      <th>application_type_Individual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20000.0</td>\n",
       "      <td>0.1485</td>\n",
       "      <td>0.0</td>\n",
       "      <td>690.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.845</td>\n",
       "      <td>12.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5356.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>102.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>37491.0</td>\n",
       "      <td>72</td>\n",
       "      <td>0.1245</td>\n",
       "      <td>35.975100</td>\n",
       "      <td>4889.01</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20000.0</td>\n",
       "      <td>0.1577</td>\n",
       "      <td>0.0</td>\n",
       "      <td>680.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.366</td>\n",
       "      <td>63.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2759.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>137.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>39.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>92.1</td>\n",
       "      <td>264.0</td>\n",
       "      <td>63456.0</td>\n",
       "      <td>60</td>\n",
       "      <td>0.2221</td>\n",
       "      <td>31.076057</td>\n",
       "      <td>1780.59</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-46.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>0.1797</td>\n",
       "      <td>0.0</td>\n",
       "      <td>685.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.780</td>\n",
       "      <td>28.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>20578.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>176.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71518.0</td>\n",
       "      <td>24</td>\n",
       "      <td>0.3570</td>\n",
       "      <td>21.901647</td>\n",
       "      <td>-4441.80</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>20000.0</td>\n",
       "      <td>0.0849</td>\n",
       "      <td>1.0</td>\n",
       "      <td>705.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.057</td>\n",
       "      <td>15.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>17700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>93.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27937.0</td>\n",
       "      <td>120</td>\n",
       "      <td>0.1761</td>\n",
       "      <td>34.119870</td>\n",
       "      <td>1538.51</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-81.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>0.0649</td>\n",
       "      <td>0.0</td>\n",
       "      <td>685.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.345</td>\n",
       "      <td>23.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1997.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>95.7</td>\n",
       "      <td>8341.0</td>\n",
       "      <td>27957.0</td>\n",
       "      <td>72</td>\n",
       "      <td>0.1307</td>\n",
       "      <td>35.891570</td>\n",
       "      <td>998.97</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   funded_amnt  int_rate  delinq_2yrs  fico_range_low  inq_last_6mths  \\\n",
       "0      20000.0    0.1485          0.0           690.0             0.0   \n",
       "1      20000.0    0.1577          0.0           680.0             0.0   \n",
       "2      10000.0    0.1797          0.0           685.0             0.0   \n",
       "3      20000.0    0.0849          1.0           705.0             0.0   \n",
       "4      10000.0    0.0649          0.0           685.0             1.0   \n",
       "\n",
       "   mths_since_last_delinq  open_acc  pub_rec  pub_rec_bankruptcies  \\\n",
       "0                     0.0       8.0      0.0                   0.0   \n",
       "1                    38.0      24.0      0.0                   0.0   \n",
       "2                     0.0      14.0      0.0                   0.0   \n",
       "3                     3.0       8.0      0.0                   0.0   \n",
       "4                     0.0      14.0      1.0                   1.0   \n",
       "\n",
       "   revol_util  total_acc  collections_12_mths_ex_med  acc_now_delinq  \\\n",
       "0       0.845       12.0                         0.0             0.0   \n",
       "1       0.366       63.0                         0.0             0.0   \n",
       "2       0.780       28.0                         0.0             0.0   \n",
       "3       0.057       15.0                         0.0             0.0   \n",
       "4       0.345       23.0                         0.0             0.0   \n",
       "\n",
       "   chargeoff_within_12_mths  acc_open_past_24mths  avg_cur_bal  delinq_amnt  \\\n",
       "0                       0.0                   3.0       5356.0          0.0   \n",
       "1                       0.0                   9.0       2759.0          0.0   \n",
       "2                       0.0                   4.0      20578.0          0.0   \n",
       "3                       0.0                   4.0      17700.0          0.0   \n",
       "4                       0.0                   7.0       1997.0          0.0   \n",
       "\n",
       "   mo_sin_old_rev_tl_op  mo_sin_rcnt_rev_tl_op  mo_sin_rcnt_tl  mort_acc  \\\n",
       "0                 102.0                   16.0            10.0       0.0   \n",
       "1                 137.0                    2.0             2.0       1.0   \n",
       "2                 176.0                    7.0             6.0       6.0   \n",
       "3                  55.0                   32.0            13.0       3.0   \n",
       "4                 129.0                    1.0             1.0       1.0   \n",
       "\n",
       "   mths_since_recent_inq  mths_since_recent_revol_delinq  \\\n",
       "0                    9.0                             0.0   \n",
       "1                    9.0                             0.0   \n",
       "2                    6.0                             0.0   \n",
       "3                    8.0                             0.0   \n",
       "4                    1.0                             0.0   \n",
       "\n",
       "   num_accts_ever_120_pd  num_il_tl  num_tl_120dpd_2m  pct_tl_nvr_dlq  \\\n",
       "0                    0.0        1.0               0.0           100.0   \n",
       "1                    5.0       39.0               0.0            92.1   \n",
       "2                    0.0        6.0               0.0           100.0   \n",
       "3                    1.0        9.0               0.0            93.3   \n",
       "4                    0.0        3.0               0.0            95.7   \n",
       "\n",
       "   tot_coll_amt  total_bal_ex_mort  emp_length_months  dti_dec  \\\n",
       "0           0.0            37491.0                 72   0.1245   \n",
       "1         264.0            63456.0                 60   0.2221   \n",
       "2           0.0            71518.0                 24   0.3570   \n",
       "3           0.0            27937.0                120   0.1761   \n",
       "4        8341.0            27957.0                 72   0.1307   \n",
       "\n",
       "   total_payments  net_gain  debt_consolidation  consumer_credit  other  \\\n",
       "0       35.975100   4889.01                   1                0      0   \n",
       "1       31.076057   1780.59                   0                1      0   \n",
       "2       21.901647  -4441.80                   1                0      0   \n",
       "3       34.119870   1538.51                   0                1      0   \n",
       "4       35.891570    998.97                   1                0      0   \n",
       "\n",
       "   countdown_zero_delinq  countdown_zero_revol_delinq  term_ 36 months  \\\n",
       "0                    0.0                          0.0                1   \n",
       "1                  -46.0                          0.0                1   \n",
       "2                    0.0                          0.0                0   \n",
       "3                  -81.0                          0.0                1   \n",
       "4                    0.0                          0.0                1   \n",
       "\n",
       "   grade_A  grade_B  grade_C  grade_D  grade_E  grade_F  grade_G  \\\n",
       "0        0        0        1        0        0        0        0   \n",
       "1        0        0        0        1        0        0        0   \n",
       "2        0        0        0        1        0        0        0   \n",
       "3        0        1        0        0        0        0        0   \n",
       "4        1        0        0        0        0        0        0   \n",
       "\n",
       "   home_ownership_ANY  home_ownership_MORTGAGE  home_ownership_OWN  \\\n",
       "0                   0                        0                   0   \n",
       "1                   0                        0                   0   \n",
       "2                   0                        1                   0   \n",
       "3                   0                        1                   0   \n",
       "4                   0                        0                   0   \n",
       "\n",
       "   verification_status_Not Verified  verification_status_Source Verified  \\\n",
       "0                                 1                                    0   \n",
       "1                                 1                                    0   \n",
       "2                                 1                                    0   \n",
       "3                                 1                                    0   \n",
       "4                                 1                                    0   \n",
       "\n",
       "   verification_status_Verified  paid  application_type_Individual  \n",
       "0                             0     1                            1  \n",
       "1                             0     1                            1  \n",
       "2                             0     0                            1  \n",
       "3                             0     1                            1  \n",
       "4                             0     1                            1  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# copy the data, so as to preserve the original dataset \n",
    "ml_data = data.copy()\n",
    "\n",
    "# # use list_cor to remove the correlated columns from the data frame \n",
    "\n",
    "ml_data = eda_functions.subset_data(ml_data, list_cor)\n",
    "\n",
    "ml_data.head()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "401819db",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 263 ms, sys: 44.3 ms, total: 307 ms\n",
      "Wall time: 306 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# Y is the target variable, I.e. whether or not a loan has been paid. A \"1\" in that column indicates a paid loan\n",
    "# while a 0 indicates a defaulted loan. \n",
    "\n",
    "# use the copy feature so we have the original data frame for later analysis (if needed)\n",
    "X =  ml_data.copy()\n",
    "y = ml_data['paid']\n",
    "\n",
    "# split into test and training data sets\n",
    "\n",
    "# split data into training and testing data sets 80:20 respectively\n",
    "# added random state for reproducibility of results \n",
    "# stratify parameter ensures that the test and train datasets maintain the same ratio of defaults vs. paid\n",
    "# loans as observed in the original dataset \n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state = 1, stratify = y)\n",
    "\n",
    "# get a copy of the test portfolio to assess model's financial performance \n",
    "test_portfolio = X_test.copy()\n",
    "\n",
    "# reset indices \n",
    "\n",
    "test_portfolio = test_portfolio.reset_index(drop=True)\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "942e50e3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# columns to be dropped from X \n",
    "columns = ['total_payments', 'net_gain', 'paid']\n",
    "\n",
    "\n",
    "# drop columns \n",
    "X.drop(columns, inplace=True, axis=1)\n",
    "X_test.drop(columns, inplace=True, axis=1)\n",
    "X_train.drop(columns, inplace=True, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a98f3aa9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>total_loan_value</th>\n",
       "      <th>total_net_gains</th>\n",
       "      <th>avg_interest rate</th>\n",
       "      <th>avg_loan_amount</th>\n",
       "      <th>avg_gains($)</th>\n",
       "      <th>avg_gains(%)</th>\n",
       "      <th>annualized_return</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1180149275.0</td>\n",
       "      <td>69884704.400058</td>\n",
       "      <td>12.54789</td>\n",
       "      <td>15283.740093</td>\n",
       "      <td>905.054709</td>\n",
       "      <td>5.921683</td>\n",
       "      <td>1.920887</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  total_loan_value  total_net_gains avg_interest rate avg_loan_amount  \\\n",
       "0     1180149275.0  69884704.400058          12.54789    15283.740093   \n",
       "\n",
       "  avg_gains($) avg_gains(%) annualized_return  \n",
       "0   905.054709     5.921683          1.920887  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# for reference, calculate the performance of the test data set \n",
    "\n",
    "portfolio_performance = eda_functions.calc_performance(test_portfolio)\n",
    "portfolio_performance.head()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2d1f7386",
   "metadata": {},
   "source": [
    "For reference the values for the entire data set are:\n",
    "* Interest Rate: 12.553\n",
    "* Loan Amount: 15,307.36\n",
    "* Annualized Return: 1.9374 "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "63a9df3b",
   "metadata": {},
   "source": [
    "#### Machine learning process \n",
    "\n",
    "This machine learning component of this project will be conducted as follows:\n",
    "* Fit a model using XG Boost and parameters determined earlier & generate a classification report\n",
    "* Using Stratified Cross Validation to validate our key metric (F1 score) \n",
    "* Estimate the performance of the loan portfolio selected by the model \n",
    "* Look at feature importances to see what features contributed the most to the model's performance\n",
    "* Experiment with various thresholds to see if removing low value features improves the model (or at least makes it simpler without losing much in the way of accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7efaca5b",
   "metadata": {},
   "source": [
    "Before we run the machine learning models, we'll need to \"scale\" the data so that all the numbers are between \n",
    "1 and 0, the reasons for this are as follows:\n",
    "* Machine learning models will treat \"higher\" numbers are being more important than lower ones\n",
    "* The model just applies math to solve a problem, it doesn't undertand context, units, etc. I.e. it can't put a 714 FICO, an interest rate of 8.9%, a debt to income ratio of 0.20 and an income of 120k in the proper context, instead it would just treat the incoome as more important than FICO\n",
    "* Scaling the numbers enables the model to treat all numbers equally \n",
    "* \"Technically\" an XG Boost model shouldn't be affected by the magnitude of the numbers, it's just one of those things where it \"couldn't hurt\". "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f6a53738",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 117 ms, sys: 32.6 ms, total: 150 ms\n",
      "Wall time: 149 ms\n"
     ]
    },
    {
     "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>funded_amnt</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>delinq_2yrs</th>\n",
       "      <th>fico_range_low</th>\n",
       "      <th>inq_last_6mths</th>\n",
       "      <th>mths_since_last_delinq</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>collections_12_mths_ex_med</th>\n",
       "      <th>acc_now_delinq</th>\n",
       "      <th>chargeoff_within_12_mths</th>\n",
       "      <th>acc_open_past_24mths</th>\n",
       "      <th>avg_cur_bal</th>\n",
       "      <th>delinq_amnt</th>\n",
       "      <th>mo_sin_old_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_tl</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>mths_since_recent_inq</th>\n",
       "      <th>mths_since_recent_revol_delinq</th>\n",
       "      <th>num_accts_ever_120_pd</th>\n",
       "      <th>num_il_tl</th>\n",
       "      <th>num_tl_120dpd_2m</th>\n",
       "      <th>pct_tl_nvr_dlq</th>\n",
       "      <th>tot_coll_amt</th>\n",
       "      <th>total_bal_ex_mort</th>\n",
       "      <th>emp_length_months</th>\n",
       "      <th>dti_dec</th>\n",
       "      <th>debt_consolidation</th>\n",
       "      <th>consumer_credit</th>\n",
       "      <th>other</th>\n",
       "      <th>countdown_zero_delinq</th>\n",
       "      <th>countdown_zero_revol_delinq</th>\n",
       "      <th>term_ 36 months</th>\n",
       "      <th>grade_A</th>\n",
       "      <th>grade_B</th>\n",
       "      <th>grade_C</th>\n",
       "      <th>grade_D</th>\n",
       "      <th>grade_E</th>\n",
       "      <th>grade_F</th>\n",
       "      <th>grade_G</th>\n",
       "      <th>home_ownership_ANY</th>\n",
       "      <th>home_ownership_MORTGAGE</th>\n",
       "      <th>home_ownership_OWN</th>\n",
       "      <th>verification_status_Not Verified</th>\n",
       "      <th>verification_status_Source Verified</th>\n",
       "      <th>verification_status_Verified</th>\n",
       "      <th>application_type_Individual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.108576</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.405405</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.191011</td>\n",
       "      <td>0.011628</td>\n",
       "      <td>0.090909</td>\n",
       "      <td>0.125389</td>\n",
       "      <td>0.175758</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.078125</td>\n",
       "      <td>0.028275</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.241503</td>\n",
       "      <td>0.053968</td>\n",
       "      <td>0.019011</td>\n",
       "      <td>0.038462</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.121212</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.021427</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.118077</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.705882</td>\n",
       "      <td>0.474440</td>\n",
       "      <td>0.076923</td>\n",
       "      <td>0.027027</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.096591</td>\n",
       "      <td>0.123596</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.338342</td>\n",
       "      <td>0.145455</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.046875</td>\n",
       "      <td>0.006009</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.355993</td>\n",
       "      <td>0.009524</td>\n",
       "      <td>0.011407</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.096591</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.068182</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.786</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.020502</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.071731</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.091429</td>\n",
       "      <td>0.091429</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.017647</td>\n",
       "      <td>0.187579</td>\n",
       "      <td>0.102564</td>\n",
       "      <td>0.108108</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.130682</td>\n",
       "      <td>0.078652</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.138860</td>\n",
       "      <td>0.024242</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.046875</td>\n",
       "      <td>0.001909</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.059034</td>\n",
       "      <td>0.015873</td>\n",
       "      <td>0.019011</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.56</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.102564</td>\n",
       "      <td>0.030303</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.500</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.004340</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.020865</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.125714</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.386988</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.702703</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.067416</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.066321</td>\n",
       "      <td>0.084848</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.046875</td>\n",
       "      <td>0.065681</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.314848</td>\n",
       "      <td>0.057143</td>\n",
       "      <td>0.019011</td>\n",
       "      <td>0.153846</td>\n",
       "      <td>0.44</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.007576</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000867</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.009615</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.111765</td>\n",
       "      <td>0.262357</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.540541</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.056180</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.126943</td>\n",
       "      <td>0.024242</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.046875</td>\n",
       "      <td>0.016732</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.320215</td>\n",
       "      <td>0.136508</td>\n",
       "      <td>0.030418</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.030303</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.023786</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.227692</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   funded_amnt  int_rate  delinq_2yrs  fico_range_low  inq_last_6mths  \\\n",
       "0     0.264706  0.108576     0.000000        0.405405        0.000000   \n",
       "1     0.705882  0.474440     0.076923        0.027027        0.333333   \n",
       "2     0.017647  0.187579     0.102564        0.108108        0.000000   \n",
       "3     0.264706  0.386988     0.000000        0.702703        0.000000   \n",
       "4     0.111765  0.262357     0.000000        0.540541        0.000000   \n",
       "\n",
       "   mths_since_last_delinq  open_acc   pub_rec  pub_rec_bankruptcies  \\\n",
       "0                0.000000  0.191011  0.011628              0.090909   \n",
       "1                0.096591  0.123596  0.000000              0.000000   \n",
       "2                0.130682  0.078652  0.000000              0.000000   \n",
       "3                0.000000  0.067416  0.000000              0.000000   \n",
       "4                0.000000  0.056180  0.000000              0.000000   \n",
       "\n",
       "   revol_util  total_acc  collections_12_mths_ex_med  acc_now_delinq  \\\n",
       "0    0.125389   0.175758                         0.0             0.0   \n",
       "1    0.338342   0.145455                         0.0             0.0   \n",
       "2    0.138860   0.024242                         0.0             0.0   \n",
       "3    0.066321   0.084848                         0.0             0.0   \n",
       "4    0.126943   0.024242                         0.0             0.0   \n",
       "\n",
       "   chargeoff_within_12_mths  acc_open_past_24mths  avg_cur_bal  delinq_amnt  \\\n",
       "0                       0.0              0.078125     0.028275          0.0   \n",
       "1                       0.0              0.046875     0.006009          0.0   \n",
       "2                       0.0              0.046875     0.001909          0.0   \n",
       "3                       0.0              0.046875     0.065681          0.0   \n",
       "4                       0.0              0.046875     0.016732          0.0   \n",
       "\n",
       "   mo_sin_old_rev_tl_op  mo_sin_rcnt_rev_tl_op  mo_sin_rcnt_tl  mort_acc  \\\n",
       "0              0.241503               0.053968        0.019011  0.038462   \n",
       "1              0.355993               0.009524        0.011407  0.000000   \n",
       "2              0.059034               0.015873        0.019011  0.000000   \n",
       "3              0.314848               0.057143        0.019011  0.153846   \n",
       "4              0.320215               0.136508        0.030418  0.000000   \n",
       "\n",
       "   mths_since_recent_inq  mths_since_recent_revol_delinq  \\\n",
       "0                   0.20                        0.000000   \n",
       "1                   0.12                        0.096591   \n",
       "2                   0.56                        0.000000   \n",
       "3                   0.44                        0.000000   \n",
       "4                   0.32                        0.000000   \n",
       "\n",
       "   num_accts_ever_120_pd  num_il_tl  num_tl_120dpd_2m  pct_tl_nvr_dlq  \\\n",
       "0               0.000000   0.121212               0.0           1.000   \n",
       "1               0.000000   0.068182               0.0           0.786   \n",
       "2               0.102564   0.030303               0.0           0.500   \n",
       "3               0.000000   0.007576               0.0           1.000   \n",
       "4               0.000000   0.030303               0.0           1.000   \n",
       "\n",
       "   tot_coll_amt  total_bal_ex_mort  emp_length_months   dti_dec  \\\n",
       "0           0.0           0.021427           1.000000  0.118077   \n",
       "1           0.0           0.020502           0.111111  0.071731   \n",
       "2           0.0           0.004340           0.888889  0.020865   \n",
       "3           0.0           0.000867           0.000000  0.009615   \n",
       "4           0.0           0.023786           1.000000  0.227692   \n",
       "\n",
       "   debt_consolidation  consumer_credit  other  countdown_zero_delinq  \\\n",
       "0                 1.0              0.0    0.0               0.474286   \n",
       "1                 0.0              1.0    0.0               0.091429   \n",
       "2                 0.0              1.0    0.0               0.125714   \n",
       "3                 0.0              0.0    1.0               0.474286   \n",
       "4                 1.0              0.0    0.0               0.474286   \n",
       "\n",
       "   countdown_zero_revol_delinq  term_ 36 months  grade_A  grade_B  grade_C  \\\n",
       "0                     0.474286              0.0      1.0      0.0      0.0   \n",
       "1                     0.091429              1.0      0.0      0.0      0.0   \n",
       "2                     0.474286              1.0      0.0      1.0      0.0   \n",
       "3                     0.474286              1.0      0.0      0.0      1.0   \n",
       "4                     0.474286              1.0      0.0      1.0      0.0   \n",
       "\n",
       "   grade_D  grade_E  grade_F  grade_G  home_ownership_ANY  \\\n",
       "0      0.0      0.0      0.0      0.0                 0.0   \n",
       "1      1.0      0.0      0.0      0.0                 0.0   \n",
       "2      0.0      0.0      0.0      0.0                 0.0   \n",
       "3      0.0      0.0      0.0      0.0                 0.0   \n",
       "4      0.0      0.0      0.0      0.0                 0.0   \n",
       "\n",
       "   home_ownership_MORTGAGE  home_ownership_OWN  \\\n",
       "0                      1.0                 0.0   \n",
       "1                      0.0                 0.0   \n",
       "2                      0.0                 0.0   \n",
       "3                      1.0                 0.0   \n",
       "4                      0.0                 0.0   \n",
       "\n",
       "   verification_status_Not Verified  verification_status_Source Verified  \\\n",
       "0                               1.0                                  0.0   \n",
       "1                               1.0                                  0.0   \n",
       "2                               0.0                                  1.0   \n",
       "3                               0.0                                  1.0   \n",
       "4                               0.0                                  0.0   \n",
       "\n",
       "   verification_status_Verified  application_type_Individual  \n",
       "0                           0.0                          1.0  \n",
       "1                           0.0                          1.0  \n",
       "2                           0.0                          1.0  \n",
       "3                           0.0                          1.0  \n",
       "4                           1.0                          1.0  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# create scaler \n",
    "scaler = MinMaxScaler()\n",
    "\n",
    "# transform data - the output is a numpy data frame so we have to turn it back into a pandas data frame\n",
    "# and add the column names back \n",
    "\n",
    "X_train = pd.DataFrame(scaler.fit_transform(X_train), columns = X_train.columns)\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f33ab3a7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 53.8 ms, sys: 177 µs, total: 54 ms\n",
      "Wall time: 52 ms\n"
     ]
    },
    {
     "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>funded_amnt</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>delinq_2yrs</th>\n",
       "      <th>fico_range_low</th>\n",
       "      <th>inq_last_6mths</th>\n",
       "      <th>mths_since_last_delinq</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>collections_12_mths_ex_med</th>\n",
       "      <th>acc_now_delinq</th>\n",
       "      <th>chargeoff_within_12_mths</th>\n",
       "      <th>acc_open_past_24mths</th>\n",
       "      <th>avg_cur_bal</th>\n",
       "      <th>delinq_amnt</th>\n",
       "      <th>mo_sin_old_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_tl</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>mths_since_recent_inq</th>\n",
       "      <th>mths_since_recent_revol_delinq</th>\n",
       "      <th>num_accts_ever_120_pd</th>\n",
       "      <th>num_il_tl</th>\n",
       "      <th>num_tl_120dpd_2m</th>\n",
       "      <th>pct_tl_nvr_dlq</th>\n",
       "      <th>tot_coll_amt</th>\n",
       "      <th>total_bal_ex_mort</th>\n",
       "      <th>emp_length_months</th>\n",
       "      <th>dti_dec</th>\n",
       "      <th>debt_consolidation</th>\n",
       "      <th>consumer_credit</th>\n",
       "      <th>other</th>\n",
       "      <th>countdown_zero_delinq</th>\n",
       "      <th>countdown_zero_revol_delinq</th>\n",
       "      <th>term_ 36 months</th>\n",
       "      <th>grade_A</th>\n",
       "      <th>grade_B</th>\n",
       "      <th>grade_C</th>\n",
       "      <th>grade_D</th>\n",
       "      <th>grade_E</th>\n",
       "      <th>grade_F</th>\n",
       "      <th>grade_G</th>\n",
       "      <th>home_ownership_ANY</th>\n",
       "      <th>home_ownership_MORTGAGE</th>\n",
       "      <th>home_ownership_OWN</th>\n",
       "      <th>verification_status_Not Verified</th>\n",
       "      <th>verification_status_Source Verified</th>\n",
       "      <th>verification_status_Verified</th>\n",
       "      <th>application_type_Individual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.794118</td>\n",
       "      <td>0.108576</td>\n",
       "      <td>0.1250</td>\n",
       "      <td>0.189189</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.038961</td>\n",
       "      <td>0.256410</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.310365</td>\n",
       "      <td>0.268657</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.170213</td>\n",
       "      <td>0.033942</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.438849</td>\n",
       "      <td>0.024691</td>\n",
       "      <td>0.049383</td>\n",
       "      <td>0.085106</td>\n",
       "      <td>0.041667</td>\n",
       "      <td>0.110390</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.942857</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.035320</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.045778</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.032680</td>\n",
       "      <td>0.104575</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.067596</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.216216</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.415584</td>\n",
       "      <td>0.051282</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.164179</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.025622</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.453237</td>\n",
       "      <td>0.092593</td>\n",
       "      <td>0.172840</td>\n",
       "      <td>0.063830</td>\n",
       "      <td>0.416667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.032258</td>\n",
       "      <td>0.010101</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.950857</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.024378</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.023178</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.411765</td>\n",
       "      <td>0.542484</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.411765</td>\n",
       "      <td>0.294466</td>\n",
       "      <td>0.0625</td>\n",
       "      <td>0.162162</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.123377</td>\n",
       "      <td>0.525641</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.113241</td>\n",
       "      <td>0.402985</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.361702</td>\n",
       "      <td>0.010584</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.294964</td>\n",
       "      <td>0.009259</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.085106</td>\n",
       "      <td>0.208333</td>\n",
       "      <td>0.123377</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.030303</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.940571</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.007909</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.035398</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.117647</td>\n",
       "      <td>0.117647</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.248529</td>\n",
       "      <td>0.788762</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.054054</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.175325</td>\n",
       "      <td>0.064103</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.128820</td>\n",
       "      <td>0.067164</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.148936</td>\n",
       "      <td>0.002150</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.167266</td>\n",
       "      <td>0.009259</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.175325</td>\n",
       "      <td>0.064516</td>\n",
       "      <td>0.050505</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.809143</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.002455</td>\n",
       "      <td>0.222222</td>\n",
       "      <td>0.038157</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.169935</td>\n",
       "      <td>0.169935</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.117647</td>\n",
       "      <td>0.281791</td>\n",
       "      <td>0.0625</td>\n",
       "      <td>0.351351</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.090909</td>\n",
       "      <td>0.153846</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.137208</td>\n",
       "      <td>0.134328</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.234043</td>\n",
       "      <td>0.014446</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.327338</td>\n",
       "      <td>0.006173</td>\n",
       "      <td>0.012346</td>\n",
       "      <td>0.021277</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.070707</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.948571</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.011853</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.072688</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.084967</td>\n",
       "      <td>0.542484</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   funded_amnt  int_rate  delinq_2yrs  fico_range_low  inq_last_6mths  \\\n",
       "0     0.794118  0.108576       0.1250        0.189189        0.000000   \n",
       "1     1.000000  0.067596       0.0000        0.216216        0.000000   \n",
       "2     0.411765  0.294466       0.0625        0.162162        0.333333   \n",
       "3     0.248529  0.788762       0.0000        0.054054        0.000000   \n",
       "4     0.117647  0.281791       0.0625        0.351351        0.166667   \n",
       "\n",
       "   mths_since_last_delinq  open_acc  pub_rec  pub_rec_bankruptcies  \\\n",
       "0                0.038961  0.256410      0.0                   0.0   \n",
       "1                0.415584  0.051282      0.0                   0.0   \n",
       "2                0.123377  0.525641      0.0                   0.0   \n",
       "3                0.175325  0.064103      0.0                   0.0   \n",
       "4                0.090909  0.153846      0.0                   0.0   \n",
       "\n",
       "   revol_util  total_acc  collections_12_mths_ex_med  acc_now_delinq  \\\n",
       "0    0.310365   0.268657                         0.0             0.0   \n",
       "1    0.000000   0.164179                         0.0             0.0   \n",
       "2    0.113241   0.402985                         0.0             0.0   \n",
       "3    0.128820   0.067164                         0.0             0.0   \n",
       "4    0.137208   0.134328                         0.0             0.0   \n",
       "\n",
       "   chargeoff_within_12_mths  acc_open_past_24mths  avg_cur_bal  delinq_amnt  \\\n",
       "0                       0.0              0.170213     0.033942          0.0   \n",
       "1                       0.0              0.000000     0.025622          0.0   \n",
       "2                       0.0              0.361702     0.010584          0.0   \n",
       "3                       0.0              0.148936     0.002150          0.0   \n",
       "4                       0.0              0.234043     0.014446          0.0   \n",
       "\n",
       "   mo_sin_old_rev_tl_op  mo_sin_rcnt_rev_tl_op  mo_sin_rcnt_tl  mort_acc  \\\n",
       "0              0.438849               0.024691        0.049383  0.085106   \n",
       "1              0.453237               0.092593        0.172840  0.063830   \n",
       "2              0.294964               0.009259        0.018519  0.085106   \n",
       "3              0.167266               0.009259        0.018519  0.000000   \n",
       "4              0.327338               0.006173        0.012346  0.021277   \n",
       "\n",
       "   mths_since_recent_inq  mths_since_recent_revol_delinq  \\\n",
       "0               0.041667                        0.110390   \n",
       "1               0.416667                        0.000000   \n",
       "2               0.208333                        0.123377   \n",
       "3               0.791667                        0.175325   \n",
       "4               0.250000                        0.000000   \n",
       "\n",
       "   num_accts_ever_120_pd  num_il_tl  num_tl_120dpd_2m  pct_tl_nvr_dlq  \\\n",
       "0               0.000000   0.111111               0.0        0.942857   \n",
       "1               0.032258   0.010101               0.0        0.950857   \n",
       "2               0.000000   0.030303               0.0        0.940571   \n",
       "3               0.064516   0.050505               0.0        0.809143   \n",
       "4               0.000000   0.070707               0.0        0.948571   \n",
       "\n",
       "   tot_coll_amt  total_bal_ex_mort  emp_length_months   dti_dec  \\\n",
       "0           0.0           0.035320           1.000000  0.045778   \n",
       "1           0.0           0.024378           0.000000  0.023178   \n",
       "2           0.0           0.007909           0.111111  0.035398   \n",
       "3           0.0           0.002455           0.222222  0.038157   \n",
       "4           0.0           0.011853           1.000000  0.072688   \n",
       "\n",
       "   debt_consolidation  consumer_credit  other  countdown_zero_delinq  \\\n",
       "0                 1.0              0.0    0.0               0.032680   \n",
       "1                 1.0              0.0    0.0               0.411765   \n",
       "2                 1.0              0.0    0.0               0.117647   \n",
       "3                 0.0              1.0    0.0               0.169935   \n",
       "4                 1.0              0.0    0.0               0.084967   \n",
       "\n",
       "   countdown_zero_revol_delinq  term_ 36 months  grade_A  grade_B  grade_C  \\\n",
       "0                     0.104575              1.0      1.0      0.0      0.0   \n",
       "1                     0.542484              1.0      1.0      0.0      0.0   \n",
       "2                     0.117647              1.0      0.0      0.0      1.0   \n",
       "3                     0.169935              1.0      0.0      0.0      0.0   \n",
       "4                     0.542484              1.0      0.0      0.0      1.0   \n",
       "\n",
       "   grade_D  grade_E  grade_F  grade_G  home_ownership_ANY  \\\n",
       "0      0.0      0.0      0.0      0.0                 0.0   \n",
       "1      0.0      0.0      0.0      0.0                 0.0   \n",
       "2      0.0      0.0      0.0      0.0                 0.0   \n",
       "3      0.0      0.0      1.0      0.0                 0.0   \n",
       "4      0.0      0.0      0.0      0.0                 0.0   \n",
       "\n",
       "   home_ownership_MORTGAGE  home_ownership_OWN  \\\n",
       "0                      0.0                 1.0   \n",
       "1                      0.0                 0.0   \n",
       "2                      1.0                 0.0   \n",
       "3                      0.0                 0.0   \n",
       "4                      1.0                 0.0   \n",
       "\n",
       "   verification_status_Not Verified  verification_status_Source Verified  \\\n",
       "0                               0.0                                  1.0   \n",
       "1                               0.0                                  0.0   \n",
       "2                               0.0                                  0.0   \n",
       "3                               0.0                                  1.0   \n",
       "4                               0.0                                  0.0   \n",
       "\n",
       "   verification_status_Verified  application_type_Individual  \n",
       "0                           0.0                          1.0  \n",
       "1                           1.0                          1.0  \n",
       "2                           1.0                          1.0  \n",
       "3                           0.0                          1.0  \n",
       "4                           1.0                          1.0  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# scale the testing data \n",
    "\n",
    "X_test = pd.DataFrame(scaler.fit_transform(X_test), columns = X_test.columns)\n",
    "X_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c6a7a16c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>funded_amnt</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>delinq_2yrs</th>\n",
       "      <th>fico_range_low</th>\n",
       "      <th>inq_last_6mths</th>\n",
       "      <th>mths_since_last_delinq</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>collections_12_mths_ex_med</th>\n",
       "      <th>acc_now_delinq</th>\n",
       "      <th>chargeoff_within_12_mths</th>\n",
       "      <th>acc_open_past_24mths</th>\n",
       "      <th>avg_cur_bal</th>\n",
       "      <th>delinq_amnt</th>\n",
       "      <th>mo_sin_old_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_tl</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>mths_since_recent_inq</th>\n",
       "      <th>mths_since_recent_revol_delinq</th>\n",
       "      <th>num_accts_ever_120_pd</th>\n",
       "      <th>num_il_tl</th>\n",
       "      <th>num_tl_120dpd_2m</th>\n",
       "      <th>pct_tl_nvr_dlq</th>\n",
       "      <th>tot_coll_amt</th>\n",
       "      <th>total_bal_ex_mort</th>\n",
       "      <th>emp_length_months</th>\n",
       "      <th>dti_dec</th>\n",
       "      <th>debt_consolidation</th>\n",
       "      <th>consumer_credit</th>\n",
       "      <th>other</th>\n",
       "      <th>countdown_zero_delinq</th>\n",
       "      <th>countdown_zero_revol_delinq</th>\n",
       "      <th>term_ 36 months</th>\n",
       "      <th>grade_A</th>\n",
       "      <th>grade_B</th>\n",
       "      <th>grade_C</th>\n",
       "      <th>grade_D</th>\n",
       "      <th>grade_E</th>\n",
       "      <th>grade_F</th>\n",
       "      <th>grade_G</th>\n",
       "      <th>home_ownership_ANY</th>\n",
       "      <th>home_ownership_MORTGAGE</th>\n",
       "      <th>home_ownership_OWN</th>\n",
       "      <th>verification_status_Not Verified</th>\n",
       "      <th>verification_status_Source Verified</th>\n",
       "      <th>verification_status_Verified</th>\n",
       "      <th>application_type_Individual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.558824</td>\n",
       "      <td>0.402619</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.162162</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.078652</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.437824</td>\n",
       "      <td>0.048485</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.046875</td>\n",
       "      <td>0.009634</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.177102</td>\n",
       "      <td>0.049383</td>\n",
       "      <td>0.038023</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.007576</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.012833</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.032718</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.558824</td>\n",
       "      <td>0.441487</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.108108</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.215909</td>\n",
       "      <td>0.258427</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.189637</td>\n",
       "      <td>0.357576</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.140625</td>\n",
       "      <td>0.004963</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.239714</td>\n",
       "      <td>0.006173</td>\n",
       "      <td>0.007605</td>\n",
       "      <td>0.019231</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.128205</td>\n",
       "      <td>0.295455</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.921</td>\n",
       "      <td>0.000532</td>\n",
       "      <td>0.021720</td>\n",
       "      <td>0.444444</td>\n",
       "      <td>0.058366</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.211429</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.534432</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.135135</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.146067</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.404145</td>\n",
       "      <td>0.145455</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.062500</td>\n",
       "      <td>0.037016</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.309481</td>\n",
       "      <td>0.021605</td>\n",
       "      <td>0.022814</td>\n",
       "      <td>0.115385</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.045455</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.024479</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.093817</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.558824</td>\n",
       "      <td>0.133925</td>\n",
       "      <td>0.025641</td>\n",
       "      <td>0.243243</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.017045</td>\n",
       "      <td>0.078652</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.029534</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.062500</td>\n",
       "      <td>0.031839</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.093023</td>\n",
       "      <td>0.098765</td>\n",
       "      <td>0.049430</td>\n",
       "      <td>0.057692</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.025641</td>\n",
       "      <td>0.068182</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.933</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.009562</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.046278</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.011429</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.049430</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.135135</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.146067</td>\n",
       "      <td>0.011628</td>\n",
       "      <td>0.090909</td>\n",
       "      <td>0.178756</td>\n",
       "      <td>0.115152</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.109375</td>\n",
       "      <td>0.003592</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.225403</td>\n",
       "      <td>0.003086</td>\n",
       "      <td>0.003802</td>\n",
       "      <td>0.019231</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.022727</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.957</td>\n",
       "      <td>0.016794</td>\n",
       "      <td>0.009569</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.034347</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>0.474286</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   funded_amnt  int_rate  delinq_2yrs  fico_range_low  inq_last_6mths  \\\n",
       "0     0.558824  0.402619     0.000000        0.162162        0.000000   \n",
       "1     0.558824  0.441487     0.000000        0.108108        0.000000   \n",
       "2     0.264706  0.534432     0.000000        0.135135        0.000000   \n",
       "3     0.558824  0.133925     0.025641        0.243243        0.000000   \n",
       "4     0.264706  0.049430     0.000000        0.135135        0.166667   \n",
       "\n",
       "   mths_since_last_delinq  open_acc   pub_rec  pub_rec_bankruptcies  \\\n",
       "0                0.000000  0.078652  0.000000              0.000000   \n",
       "1                0.215909  0.258427  0.000000              0.000000   \n",
       "2                0.000000  0.146067  0.000000              0.000000   \n",
       "3                0.017045  0.078652  0.000000              0.000000   \n",
       "4                0.000000  0.146067  0.011628              0.090909   \n",
       "\n",
       "   revol_util  total_acc  collections_12_mths_ex_med  acc_now_delinq  \\\n",
       "0    0.437824   0.048485                         0.0             0.0   \n",
       "1    0.189637   0.357576                         0.0             0.0   \n",
       "2    0.404145   0.145455                         0.0             0.0   \n",
       "3    0.029534   0.066667                         0.0             0.0   \n",
       "4    0.178756   0.115152                         0.0             0.0   \n",
       "\n",
       "   chargeoff_within_12_mths  acc_open_past_24mths  avg_cur_bal  delinq_amnt  \\\n",
       "0                       0.0              0.046875     0.009634          0.0   \n",
       "1                       0.0              0.140625     0.004963          0.0   \n",
       "2                       0.0              0.062500     0.037016          0.0   \n",
       "3                       0.0              0.062500     0.031839          0.0   \n",
       "4                       0.0              0.109375     0.003592          0.0   \n",
       "\n",
       "   mo_sin_old_rev_tl_op  mo_sin_rcnt_rev_tl_op  mo_sin_rcnt_tl  mort_acc  \\\n",
       "0              0.177102               0.049383        0.038023  0.000000   \n",
       "1              0.239714               0.006173        0.007605  0.019231   \n",
       "2              0.309481               0.021605        0.022814  0.115385   \n",
       "3              0.093023               0.098765        0.049430  0.057692   \n",
       "4              0.225403               0.003086        0.003802  0.019231   \n",
       "\n",
       "   mths_since_recent_inq  mths_since_recent_revol_delinq  \\\n",
       "0                   0.36                             0.0   \n",
       "1                   0.36                             0.0   \n",
       "2                   0.24                             0.0   \n",
       "3                   0.32                             0.0   \n",
       "4                   0.04                             0.0   \n",
       "\n",
       "   num_accts_ever_120_pd  num_il_tl  num_tl_120dpd_2m  pct_tl_nvr_dlq  \\\n",
       "0               0.000000   0.007576               0.0           1.000   \n",
       "1               0.128205   0.295455               0.0           0.921   \n",
       "2               0.000000   0.045455               0.0           1.000   \n",
       "3               0.025641   0.068182               0.0           0.933   \n",
       "4               0.000000   0.022727               0.0           0.957   \n",
       "\n",
       "   tot_coll_amt  total_bal_ex_mort  emp_length_months   dti_dec  \\\n",
       "0      0.000000           0.012833           0.555556  0.032718   \n",
       "1      0.000532           0.021720           0.444444  0.058366   \n",
       "2      0.000000           0.024479           0.111111  0.093817   \n",
       "3      0.000000           0.009562           1.000000  0.046278   \n",
       "4      0.016794           0.009569           0.555556  0.034347   \n",
       "\n",
       "   debt_consolidation  consumer_credit  other  countdown_zero_delinq  \\\n",
       "0                 1.0              0.0    0.0               0.474286   \n",
       "1                 0.0              1.0    0.0               0.211429   \n",
       "2                 1.0              0.0    0.0               0.474286   \n",
       "3                 0.0              1.0    0.0               0.011429   \n",
       "4                 1.0              0.0    0.0               0.474286   \n",
       "\n",
       "   countdown_zero_revol_delinq  term_ 36 months  grade_A  grade_B  grade_C  \\\n",
       "0                     0.474286              1.0      0.0      0.0      1.0   \n",
       "1                     0.474286              1.0      0.0      0.0      0.0   \n",
       "2                     0.474286              0.0      0.0      0.0      0.0   \n",
       "3                     0.474286              1.0      0.0      1.0      0.0   \n",
       "4                     0.474286              1.0      1.0      0.0      0.0   \n",
       "\n",
       "   grade_D  grade_E  grade_F  grade_G  home_ownership_ANY  \\\n",
       "0      0.0      0.0      0.0      0.0                 0.0   \n",
       "1      1.0      0.0      0.0      0.0                 0.0   \n",
       "2      1.0      0.0      0.0      0.0                 0.0   \n",
       "3      0.0      0.0      0.0      0.0                 0.0   \n",
       "4      0.0      0.0      0.0      0.0                 0.0   \n",
       "\n",
       "   home_ownership_MORTGAGE  home_ownership_OWN  \\\n",
       "0                      0.0                 0.0   \n",
       "1                      0.0                 0.0   \n",
       "2                      1.0                 0.0   \n",
       "3                      1.0                 0.0   \n",
       "4                      0.0                 0.0   \n",
       "\n",
       "   verification_status_Not Verified  verification_status_Source Verified  \\\n",
       "0                               1.0                                  0.0   \n",
       "1                               1.0                                  0.0   \n",
       "2                               1.0                                  0.0   \n",
       "3                               1.0                                  0.0   \n",
       "4                               1.0                                  0.0   \n",
       "\n",
       "   verification_status_Verified  application_type_Individual  \n",
       "0                           0.0                          1.0  \n",
       "1                           0.0                          1.0  \n",
       "2                           0.0                          1.0  \n",
       "3                           0.0                          1.0  \n",
       "4                           0.0                          1.0  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# scale the dataset with just the features (X) but without separating it into test or training datasets\n",
    "# The reason for this doing is that the cross-validation exercise we'll run to validate our initial results \n",
    "# requires the entire dataset, not just a split of it \n",
    "\n",
    "# scale the testing data \n",
    "\n",
    "X_scale = X.copy()\n",
    "\n",
    "X_scale = pd.DataFrame(scaler.fit_transform(X_scale), columns = X_scale.columns)\n",
    "X_scale.head()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "db9f3a73",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 32 µs, sys: 3 µs, total: 35 µs\n",
      "Wall time: 38.4 µs\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# the parameters were determined via a grid search \n",
    "# gpu_id parameter enables GPU acceleration via CUDA \n",
    "# the scale_pos_weight parameter adjusts for the fact that good loans outnumber the bad loans by nearly 4:1\n",
    "# The number is calulated by dividing the bad loans (smaller class) by the good loans (larger class)\n",
    "\n",
    "# define classifier\n",
    "\n",
    "classifier = xgb.XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "              colsample_bynode=1, colsample_bytree=0.8,\n",
    "              enable_categorical=False, gamma=1.5, gpu_id=0,\n",
    "              importance_type=None, interaction_constraints='',\n",
    "              learning_rate=0.02,objective='reg:logistic',  max_delta_step=0, max_depth=5,\n",
    "              min_child_weight=1, monotone_constraints='()',\n",
    "              n_estimators=600, n_jobs=1, nthread=1, num_parallel_tree=1,\n",
    "              predictor='auto', random_state=0, reg_alpha=0, reg_lambda=1,\n",
    "              scale_pos_weight=0.237, eval_metric='auc', subsample=0.6,\n",
    "              tree_method='gpu_hist', validate_parameters=1, verbosity=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c8f3d5ea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.32      0.64      0.43     14806\n",
      "           1       0.89      0.68      0.77     62410\n",
      "\n",
      "    accuracy                           0.67     77216\n",
      "   macro avg       0.60      0.66      0.60     77216\n",
      "weighted avg       0.78      0.67      0.70     77216\n",
      "\n",
      "CPU times: user 5.07 s, sys: 112 ms, total: 5.18 s\n",
      "Wall time: 5.21 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# fit a model on the training data, test it and then generate a classification repo\n",
    "\n",
    "eval_set = [(X_test, y_test)]\n",
    "\n",
    "# the early stopping arounds, eval_metric and eval set parameters are so that the model will stop running once \n",
    "# error rates stop improving after 10 rounds. This is done to prevent overfitting. Changing the parameter \"verbose\" \n",
    "# true would show the outcome of each training epoch. I turned it off as it's a lot of data/to keep the \n",
    "# notebook cleaner \n",
    "model = classifier.fit(X_train, y_train, early_stopping_rounds=10, eval_metric=\"logloss\",\n",
    "                      eval_set=eval_set, verbose=False)\n",
    "predictions = model.predict(X_test)\n",
    "classification = classification_report(y_test, predictions)\n",
    "print(classification)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a149b01",
   "metadata": {},
   "source": [
    "#### Explaining the Report\n",
    "\n",
    "* From a pure machine learning accuracy perspective the numbers are good, but not great, primarily because the high precision score came at the cost of a significant number of false negatives.\n",
    "* The above being said, given the high cost of bad loans, the model does deliver excellent financial results, as will be explored more below. \n",
    "* Precision in this case roughly equates to the default rate if Lending Club only originated loans the model identifies as good, meaning of the loans the model identified as \"good\", 89% were good, which means the default rate was about 11%.  \n",
    "* The recall score of 0.68 for \"1\" or the positive class, means that the number of loans the model classified as good equals 68% of the total number of good loans\n",
    "* The f1-score is a \"harmonic mean\" between precision and recall, giving us a balance of both measures\n",
    "* Accuracy is \"good\" but not a great measure for this type of problem as it's an imbalanced dataset, so we're unlikely to have high accuracy for both types of loans, however, we only need great accuracy for one of the classes for the model to be successful. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6148f839",
   "metadata": {},
   "outputs": [],
   "source": [
    "# generate ROC curve to check for overfitting =2.\n",
    "\n",
    "# plot AUC curve to see relative performance vs. a no skill\n",
    "# or guessing line \n",
    "\n",
    "# AUC for testing data\n",
    "\n",
    "# predict probabilities for the paid off loans \n",
    "yhat = model.predict_proba(X_test)\n",
    "\n",
    "# get the probability values for the paid off loans \n",
    "yhat_positive = yhat[:, 1]\n",
    "\n",
    "\n",
    "# calculate inputs for the test data roc curve\n",
    "fpr_test, tpr_test, thresholds = metrics.roc_curve(y_test, yhat_positive)\n",
    "\n",
    "\n",
    "# AUC for training data\n",
    "\n",
    "# predict probabilities for the paid off loans \n",
    "yhat_train = model.predict_proba(X_train)\n",
    "\n",
    "# get the probability values for the paid off loans \n",
    "yhat_positive_train = yhat_train[:, 1]\n",
    "\n",
    "# calculate inputs for the roc curve - training data\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_train, yhat_positive_train)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0d4ac304",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot roc curve\n",
    "plt.plot(fpr_test, tpr_test, marker='.', label='logistic - test data') \n",
    "plt.plot(fpr, tpr, marker='.', label='Logistic - training data') \n",
    "plt.plot([0, 1], [0, 1], linestyle='--', label='No Skill')\n",
    "\n",
    "# axis labels\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "# show the legend\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f5889cfb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ROC Test Data</th>\n",
       "      <th>ROC Training Data</th>\n",
       "      <th>Diff in ROC Curves</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.722</td>\n",
       "      <td>0.732</td>\n",
       "      <td>0.011</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ROC Test Data  ROC Training Data  Diff in ROC Curves\n",
       "0          0.722              0.732               0.011"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# display the auc scores and calculate the delta between the AUC for the training vs. test data\n",
    "\n",
    "# calculate and print AUROC\n",
    "roc_auc = metrics.roc_auc_score(y_test, yhat_positive)\n",
    "\n",
    "# calculate and print AUROC\n",
    "roc_auc_train = metrics.roc_auc_score(y_train, yhat_positive_train)\n",
    "\n",
    "# calculate the difference in ROC curves \n",
    "roc_diff = abs(roc_auc_train - roc_auc)\n",
    "\n",
    "roc_dict = {'ROC Test Data': roc_auc,\n",
    "            'ROC Training Data': roc_auc_train,\n",
    "            'Diff in ROC Curves': roc_diff}\n",
    "\n",
    "roc_df = pd.DataFrame(roc_dict, index=[0])\n",
    "\n",
    "roc_df = roc_df.round(3)\n",
    "\n",
    "roc_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abdf3a9a",
   "metadata": {},
   "source": [
    "The model \"has skill\" given that the the AUC scores are above the 0.5 line we could achieve via just guessing, and the small difference in the curves for testing and training data suggests that we don't have any overfitting issues. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a3c9517",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>predictions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   predictions\n",
       "0            1\n",
       "1            1\n",
       "2            1\n",
       "3            0\n",
       "4            1"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# let's grab the predictions so we can determine the economic performance of the loans selected by the model \n",
    "\n",
    "columns = ['predictions']\n",
    "\n",
    "predictions_df = pd.DataFrame(predictions, columns=columns)\n",
    "predictions_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "02d695d6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>funded_amnt</th>\n",
       "      <th>int_rate</th>\n",
       "      <th>delinq_2yrs</th>\n",
       "      <th>fico_range_low</th>\n",
       "      <th>inq_last_6mths</th>\n",
       "      <th>mths_since_last_delinq</th>\n",
       "      <th>open_acc</th>\n",
       "      <th>pub_rec</th>\n",
       "      <th>pub_rec_bankruptcies</th>\n",
       "      <th>revol_util</th>\n",
       "      <th>total_acc</th>\n",
       "      <th>collections_12_mths_ex_med</th>\n",
       "      <th>acc_now_delinq</th>\n",
       "      <th>chargeoff_within_12_mths</th>\n",
       "      <th>acc_open_past_24mths</th>\n",
       "      <th>avg_cur_bal</th>\n",
       "      <th>delinq_amnt</th>\n",
       "      <th>mo_sin_old_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_rev_tl_op</th>\n",
       "      <th>mo_sin_rcnt_tl</th>\n",
       "      <th>mort_acc</th>\n",
       "      <th>mths_since_recent_inq</th>\n",
       "      <th>mths_since_recent_revol_delinq</th>\n",
       "      <th>num_accts_ever_120_pd</th>\n",
       "      <th>num_il_tl</th>\n",
       "      <th>num_tl_120dpd_2m</th>\n",
       "      <th>pct_tl_nvr_dlq</th>\n",
       "      <th>tot_coll_amt</th>\n",
       "      <th>total_bal_ex_mort</th>\n",
       "      <th>emp_length_months</th>\n",
       "      <th>dti_dec</th>\n",
       "      <th>total_payments</th>\n",
       "      <th>net_gain</th>\n",
       "      <th>debt_consolidation</th>\n",
       "      <th>consumer_credit</th>\n",
       "      <th>other</th>\n",
       "      <th>countdown_zero_delinq</th>\n",
       "      <th>countdown_zero_revol_delinq</th>\n",
       "      <th>term_ 36 months</th>\n",
       "      <th>grade_A</th>\n",
       "      <th>grade_B</th>\n",
       "      <th>grade_C</th>\n",
       "      <th>grade_D</th>\n",
       "      <th>grade_E</th>\n",
       "      <th>grade_F</th>\n",
       "      <th>grade_G</th>\n",
       "      <th>home_ownership_ANY</th>\n",
       "      <th>home_ownership_MORTGAGE</th>\n",
       "      <th>home_ownership_OWN</th>\n",
       "      <th>verification_status_Not Verified</th>\n",
       "      <th>verification_status_Source Verified</th>\n",
       "      <th>verification_status_Verified</th>\n",
       "      <th>paid</th>\n",
       "      <th>application_type_Individual</th>\n",
       "      <th>predictions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>28000.0</td>\n",
       "      <td>0.0789</td>\n",
       "      <td>2.0</td>\n",
       "      <td>695.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.518</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>18869.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>250.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>103190.0</td>\n",
       "      <td>120</td>\n",
       "      <td>0.1742</td>\n",
       "      <td>35.973398</td>\n",
       "      <td>3512.70</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-78.0</td>\n",
       "      <td>-67.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>35000.0</td>\n",
       "      <td>0.0692</td>\n",
       "      <td>0.0</td>\n",
       "      <td>700.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>14244.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>258.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>95.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71221.0</td>\n",
       "      <td>12</td>\n",
       "      <td>0.0882</td>\n",
       "      <td>36.003003</td>\n",
       "      <td>3862.36</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-20.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15000.0</td>\n",
       "      <td>0.1229</td>\n",
       "      <td>1.0</td>\n",
       "      <td>690.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.189</td>\n",
       "      <td>58.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>5884.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>170.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>94.8</td>\n",
       "      <td>0.0</td>\n",
       "      <td>23108.0</td>\n",
       "      <td>24</td>\n",
       "      <td>0.1347</td>\n",
       "      <td>31.553908</td>\n",
       "      <td>786.42</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-65.0</td>\n",
       "      <td>-65.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9450.0</td>\n",
       "      <td>0.2399</td>\n",
       "      <td>0.0</td>\n",
       "      <td>670.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.215</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1195.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>83.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7172.0</td>\n",
       "      <td>36</td>\n",
       "      <td>0.1452</td>\n",
       "      <td>16.724771</td>\n",
       "      <td>-3914.54</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-57.0</td>\n",
       "      <td>-57.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5000.0</td>\n",
       "      <td>0.1199</td>\n",
       "      <td>1.0</td>\n",
       "      <td>725.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.229</td>\n",
       "      <td>22.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>8031.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>188.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>95.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>34629.0</td>\n",
       "      <td>120</td>\n",
       "      <td>0.2766</td>\n",
       "      <td>35.875650</td>\n",
       "      <td>957.15</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-70.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   funded_amnt  int_rate  delinq_2yrs  fico_range_low  inq_last_6mths  \\\n",
       "0      28000.0    0.0789          2.0           695.0             0.0   \n",
       "1      35000.0    0.0692          0.0           700.0             0.0   \n",
       "2      15000.0    0.1229          1.0           690.0             2.0   \n",
       "3       9450.0    0.2399          0.0           670.0             0.0   \n",
       "4       5000.0    0.1199          1.0           725.0             1.0   \n",
       "\n",
       "   mths_since_last_delinq  open_acc  pub_rec  pub_rec_bankruptcies  \\\n",
       "0                     6.0      21.0      0.0                   0.0   \n",
       "1                    64.0       5.0      0.0                   0.0   \n",
       "2                    19.0      42.0      0.0                   0.0   \n",
       "3                    27.0       6.0      0.0                   0.0   \n",
       "4                    14.0      13.0      0.0                   0.0   \n",
       "\n",
       "   revol_util  total_acc  collections_12_mths_ex_med  acc_now_delinq  \\\n",
       "0       0.518       40.0                         0.0             0.0   \n",
       "1       0.000       26.0                         0.0             0.0   \n",
       "2       0.189       58.0                         0.0             0.0   \n",
       "3       0.215       13.0                         0.0             0.0   \n",
       "4       0.229       22.0                         0.0             0.0   \n",
       "\n",
       "   chargeoff_within_12_mths  acc_open_past_24mths  avg_cur_bal  delinq_amnt  \\\n",
       "0                       0.0                   8.0      18869.0          0.0   \n",
       "1                       0.0                   0.0      14244.0          0.0   \n",
       "2                       0.0                  17.0       5884.0          0.0   \n",
       "3                       0.0                   7.0       1195.0          0.0   \n",
       "4                       0.0                  11.0       8031.0          0.0   \n",
       "\n",
       "   mo_sin_old_rev_tl_op  mo_sin_rcnt_rev_tl_op  mo_sin_rcnt_tl  mort_acc  \\\n",
       "0                 250.0                    8.0             8.0       4.0   \n",
       "1                 258.0                   30.0            28.0       3.0   \n",
       "2                 170.0                    3.0             3.0       4.0   \n",
       "3                  99.0                    3.0             3.0       0.0   \n",
       "4                 188.0                    2.0             2.0       1.0   \n",
       "\n",
       "   mths_since_recent_inq  mths_since_recent_revol_delinq  \\\n",
       "0                    1.0                            17.0   \n",
       "1                   10.0                             0.0   \n",
       "2                    5.0                            19.0   \n",
       "3                   19.0                            27.0   \n",
       "4                    6.0                             0.0   \n",
       "\n",
       "   num_accts_ever_120_pd  num_il_tl  num_tl_120dpd_2m  pct_tl_nvr_dlq  \\\n",
       "0                    0.0       11.0               0.0            95.0   \n",
       "1                    1.0        1.0               0.0            95.7   \n",
       "2                    0.0        3.0               0.0            94.8   \n",
       "3                    2.0        5.0               0.0            83.3   \n",
       "4                    0.0        7.0               0.0            95.5   \n",
       "\n",
       "   tot_coll_amt  total_bal_ex_mort  emp_length_months  dti_dec  \\\n",
       "0           0.0           103190.0                120   0.1742   \n",
       "1           0.0            71221.0                 12   0.0882   \n",
       "2           0.0            23108.0                 24   0.1347   \n",
       "3           0.0             7172.0                 36   0.1452   \n",
       "4           0.0            34629.0                120   0.2766   \n",
       "\n",
       "   total_payments  net_gain  debt_consolidation  consumer_credit  other  \\\n",
       "0       35.973398   3512.70                   1                0      0   \n",
       "1       36.003003   3862.36                   1                0      0   \n",
       "2       31.553908    786.42                   1                0      0   \n",
       "3       16.724771  -3914.54                   0                1      0   \n",
       "4       35.875650    957.15                   1                0      0   \n",
       "\n",
       "   countdown_zero_delinq  countdown_zero_revol_delinq  term_ 36 months  \\\n",
       "0                  -78.0                        -67.0                1   \n",
       "1                  -20.0                          0.0                1   \n",
       "2                  -65.0                        -65.0                1   \n",
       "3                  -57.0                        -57.0                1   \n",
       "4                  -70.0                          0.0                1   \n",
       "\n",
       "   grade_A  grade_B  grade_C  grade_D  grade_E  grade_F  grade_G  \\\n",
       "0        1        0        0        0        0        0        0   \n",
       "1        1        0        0        0        0        0        0   \n",
       "2        0        0        1        0        0        0        0   \n",
       "3        0        0        0        0        0        1        0   \n",
       "4        0        0        1        0        0        0        0   \n",
       "\n",
       "   home_ownership_ANY  home_ownership_MORTGAGE  home_ownership_OWN  \\\n",
       "0                   0                        0                   1   \n",
       "1                   0                        0                   0   \n",
       "2                   0                        1                   0   \n",
       "3                   0                        0                   0   \n",
       "4                   0                        1                   0   \n",
       "\n",
       "   verification_status_Not Verified  verification_status_Source Verified  \\\n",
       "0                                 0                                    1   \n",
       "1                                 0                                    0   \n",
       "2                                 0                                    0   \n",
       "3                                 0                                    1   \n",
       "4                                 0                                    0   \n",
       "\n",
       "   verification_status_Verified  paid  application_type_Individual  \\\n",
       "0                             0     1                            1   \n",
       "1                             1     1                            1   \n",
       "2                             1     1                            1   \n",
       "3                             0     0                            1   \n",
       "4                             1     1                            1   \n",
       "\n",
       "   predictions  \n",
       "0            1  \n",
       "1            1  \n",
       "2            1  \n",
       "3            0  \n",
       "4            1  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# make a copy of the data frame created earlier for $ performance analysis \n",
    "# we do this so that we always have a fresh copy as we evaluate different models \n",
    "\n",
    "model_one = test_portfolio.copy()\n",
    "\n",
    "# add the predictions to the data frame\n",
    "\n",
    "model_one = model_one.join(predictions_df)\n",
    "\n",
    "model_one.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0e6186c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# isolate just the loans identified by the model as \"good\"\n",
    "\n",
    "model_one_selected = model_one[(model_one['predictions'] >0)]\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ba2975bd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>total_loan_value</th>\n",
       "      <th>total_net_gains</th>\n",
       "      <th>avg_interest rate</th>\n",
       "      <th>avg_loan_amount</th>\n",
       "      <th>avg_gains($)</th>\n",
       "      <th>avg_gains(%)</th>\n",
       "      <th>annualized_return</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1180149275.0</td>\n",
       "      <td>69884704.400058</td>\n",
       "      <td>12.54789</td>\n",
       "      <td>15283.740093</td>\n",
       "      <td>905.054709</td>\n",
       "      <td>5.921683</td>\n",
       "      <td>1.920887</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  total_loan_value  total_net_gains avg_interest rate avg_loan_amount  \\\n",
       "0     1180149275.0  69884704.400058          12.54789    15283.740093   \n",
       "\n",
       "  avg_gains($) avg_gains(%) annualized_return  \n",
       "0   905.054709     5.921683          1.920887  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# performance of test portfolio for reference \n",
    "\n",
    "portfolio_performance = eda_functions.calc_performance(test_portfolio)\n",
    "portfolio_performance.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7a05eae0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>total_loan_value</th>\n",
       "      <th>total_net_gains</th>\n",
       "      <th>avg_interest rate</th>\n",
       "      <th>avg_loan_amount</th>\n",
       "      <th>avg_gains($)</th>\n",
       "      <th>avg_gains(%)</th>\n",
       "      <th>annualized_return</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>675648800.0</td>\n",
       "      <td>53682768.990018</td>\n",
       "      <td>10.044185</td>\n",
       "      <td>14140.236909</td>\n",
       "      <td>1123.493554</td>\n",
       "      <td>7.945366</td>\n",
       "      <td>2.605083</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  total_loan_value  total_net_gains avg_interest rate avg_loan_amount  \\\n",
       "0      675648800.0  53682768.990018         10.044185    14140.236909   \n",
       "\n",
       "  avg_gains($) avg_gains(%) annualized_return  \n",
       "0  1123.493554     7.945366          2.605083  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# performance of the first model \n",
    "\n",
    "model_one_performance = eda_functions.calc_performance(model_one_selected)\n",
    "model_one_performance\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d4cf7b88",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>total_loan_value</th>\n",
       "      <th>total_net_gains</th>\n",
       "      <th>avg_interest rate</th>\n",
       "      <th>avg_loan_amount</th>\n",
       "      <th>avg_gains($)</th>\n",
       "      <th>avg_gains(%)</th>\n",
       "      <th>annualized_return</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3378244000.0</td>\n",
       "      <td>268413844.950092</td>\n",
       "      <td>10.044185</td>\n",
       "      <td>14140.236909</td>\n",
       "      <td>1123.493554</td>\n",
       "      <td>7.945366</td>\n",
       "      <td>2.605083</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  total_loan_value   total_net_gains avg_interest rate avg_loan_amount  \\\n",
       "0     3378244000.0  268413844.950092         10.044185    14140.236909   \n",
       "\n",
       "  avg_gains($) avg_gains(%) annualized_return  \n",
       "0  1123.493554     7.945366          2.605083  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# let's extrapolate the total loan value and total net gains to the size of the entire dataset \n",
    "\n",
    "model_one_extrapolated = model_one_performance.copy()\n",
    "\n",
    "\n",
    "model_one_extrapolated['total_loan_value'] = (5 * model_one_extrapolated['total_loan_value'])\n",
    "model_one_extrapolated['total_net_gains'] = (5 * model_one_extrapolated['total_net_gains'])\n",
    "\n",
    "model_one_extrapolated"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4501de9",
   "metadata": {},
   "source": [
    "#### Key Take Aways\n",
    "* Compared to the performance of the loans in the testing dataset, this model would increase annualized returns by 38.5% (% change from 1.92% to 2.65%).  \n",
    "\n",
    "* If we were to extrapolate to the results on the test dataset to the overall loan portfolio, we could estimate the following: \n",
    "\n",
    "    * 11.2% default rate \n",
    "\n",
    "    * 275M in returns vs. 3.4B in capital  \n",
    "\n",
    "    * The model would reduce gross returns by $80M but require nearly 2.5B less capital to do so  \n",
    "\n",
    "* This model would outperform the proposed intervention of setting an interest cap of 13.5%, because it would deliver about 7M more in revenue, while also using about 17M less capital.  \n",
    "\n",
    "* The company would generate smaller returns in absolute terms if they used this model, 22% less in absolute terms, but their % returns would increase because they would generate those returns via deploying 45% less capital. Another way to look at is that the capital cost to generate returns from their loan portfolio would be nearly cut in half.  \n",
    "\n",
    "* This model has the highest annualized returns compared to the two interventions proposed in the EDA Notebook: \n",
    "\n",
    "    * No loans below Grade D, 2.2% annualized return \n",
    "\n",
    "    * Setting an interest rate cap of 13.5%, 2.4% annualized return \n",
    "\n",
    "* The value this model would have for the company would depend on a lot of factors: namely, are they going to optimize capital efficiency or gross returns? Also, the company could elect to make up the difference with growth and marketing. I.e., aggressively seek out additional customers who meet their new lending standards. \n",
    "\n",
    "* The model and the suggested interventions from the data analysis say the same thing: after a certain point capital efficiency drops off substantially and the company is investing a considerable sum of money for a small gain. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c48ebb83",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "47782"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# of loans selected by the model as \"good\"\n",
    "\n",
    "model_one_count = len(model_one_selected)\n",
    "model_one_count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "002c391c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.11265748608262527"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# default rate \n",
    "\n",
    "default_rate_model_one = 1 - model_one_selected['paid'].mean()\n",
    "default_rate_model_one"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b5fdada6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 Score: 75.12% (0.15%)\n",
      "CPU times: user 1min 10s, sys: 618 ms, total: 1min 11s\n",
      "Wall time: 1min 11s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# before we attempt to improve the model further, let's conduct stratified cross validation where we \n",
    "# split the entire dataset into chunks (folds) and go through several training and test cycles. We do this to \n",
    "# as it its a better measure of how the model would perform on new data \n",
    "\n",
    "\n",
    "# define evaluation procedure - using Stratified Kfold + the Shuffle parameter ensures that each \"fold\"\n",
    "# has the same ratio of good to bad loans as the dataset overall. \n",
    "cv = StratifiedKFold(n_splits=10, shuffle=True, random_state=1)\n",
    "\n",
    "# evaluate model\n",
    "#results = cross_val_score(model, X_scale, y, scoring='f1', cv=cv)\n",
    "results = cross_val_score(model, X_scale, y, scoring='f1', cv=cv)\n",
    "\n",
    "print(\"F1 Score: %.2f%% (%.2f%%)\" % (results.mean()*100, results.std()*100))\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2a8dee8e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.75254111 0.7497208  0.75138122 0.75213707 0.7508832  0.74820707\n",
      " 0.75367265 0.75228162 0.75125581 0.75033517]\n"
     ]
    }
   ],
   "source": [
    "print(results)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7904e9be",
   "metadata": {},
   "source": [
    "The F1 score from cross fold validation being slightly smaller than what we see from typical model fitting, suggests that the model would perform only slightly worse on new data than it does on the training data. It's worth noting that an earlier attempt used 33% of the data for testing and delivered an F1 score of 82, while the F1 from cross validation was nearly exactly the same as is shown above. I.e. this step allowed us to do a better job of setting the testing vs. training cut off, with the outcome being a higher performing model. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6a74d3c6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show the feature importances \n",
    "\n",
    "plot_importance(model)\n",
    "pyplot.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55078bda",
   "metadata": {},
   "source": [
    "Looking at the feature importances, the general pattern is that affordability and debt load are what are driving defaults, based on the importance of debt-to-income ratio, credit card balance and other general debt metrics. Overall, these numbers are not particularly surprising: the defaults are trending in line with the customers that traditional credit measures deem as higher risk. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4be20a69",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7716700944994609\n",
      "CPU times: user 6.82 s, sys: 39.7 ms, total: 6.86 s\n",
      "Wall time: 6.88 s\n"
     ]
    }
   ],
   "source": [
    "%%time \n",
    "\n",
    "\n",
    "model_two = xgb.XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "              colsample_bynode=1, colsample_bytree=0.8,\n",
    "              enable_categorical=False, gamma=1.5, gpu_id=0,\n",
    "              importance_type=None, interaction_constraints='',\n",
    "              learning_rate=0.02,objective='reg:logistic',  max_delta_step=0, max_depth=5,\n",
    "              min_child_weight=1, monotone_constraints='()',\n",
    "              n_estimators=600, n_jobs=1, nthread=1, num_parallel_tree=1,\n",
    "              predictor='auto', random_state=0, reg_alpha=0, reg_lambda=1,\n",
    "              scale_pos_weight=0.237, eval_metric='auc', subsample=0.6,\n",
    "              tree_method='gpu_hist', validate_parameters=1, verbosity=None)\n",
    "\n",
    "model_two.fit(X_train, y_train)\n",
    "predictions = model_two.predict(X_test)\n",
    "f1_score = f1_score(y_test, predictions)\n",
    "print(f1_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dff59329",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Thresh=0.000, n=51, F1 Score: 77.17%\n",
      "Thresh=0.000, n=51, F1 Score: 77.17%\n",
      "Thresh=0.000, n=51, F1 Score: 77.17%\n",
      "Thresh=0.002, n=48, F1 Score: 76.93%\n",
      "Thresh=0.003, n=47, F1 Score: 77.17%\n",
      "Thresh=0.003, n=46, F1 Score: 77.26%\n",
      "Thresh=0.003, n=45, F1 Score: 76.88%\n",
      "Thresh=0.003, n=44, F1 Score: 77.05%\n",
      "Thresh=0.003, n=43, F1 Score: 77.30%\n",
      "Thresh=0.004, n=42, F1 Score: 77.17%\n",
      "Thresh=0.004, n=41, F1 Score: 77.36%\n",
      "Thresh=0.004, n=40, F1 Score: 77.02%\n",
      "Thresh=0.004, n=39, F1 Score: 77.03%\n",
      "Thresh=0.004, n=38, F1 Score: 76.62%\n",
      "Thresh=0.004, n=37, F1 Score: 76.22%\n",
      "Thresh=0.004, n=36, F1 Score: 76.92%\n",
      "Thresh=0.004, n=35, F1 Score: 76.49%\n",
      "Thresh=0.004, n=34, F1 Score: 76.82%\n",
      "Thresh=0.004, n=33, F1 Score: 76.11%\n",
      "Thresh=0.004, n=32, F1 Score: 76.61%\n",
      "Thresh=0.004, n=31, F1 Score: 76.12%\n",
      "Thresh=0.004, n=30, F1 Score: 75.71%\n",
      "Thresh=0.004, n=29, F1 Score: 76.24%\n",
      "Thresh=0.004, n=28, F1 Score: 76.53%\n",
      "Thresh=0.004, n=27, F1 Score: 77.66%\n",
      "Thresh=0.004, n=26, F1 Score: 77.57%\n",
      "Thresh=0.004, n=25, F1 Score: 77.31%\n",
      "Thresh=0.004, n=24, F1 Score: 77.33%\n",
      "Thresh=0.004, n=23, F1 Score: 77.38%\n",
      "Thresh=0.005, n=22, F1 Score: 77.48%\n",
      "Thresh=0.005, n=21, F1 Score: 77.27%\n",
      "Thresh=0.005, n=20, F1 Score: 77.36%\n",
      "Thresh=0.005, n=19, F1 Score: 77.03%\n",
      "Thresh=0.005, n=18, F1 Score: 76.79%\n",
      "Thresh=0.005, n=17, F1 Score: 76.77%\n",
      "Thresh=0.006, n=16, F1 Score: 77.72%\n",
      "Thresh=0.006, n=15, F1 Score: 77.62%\n",
      "Thresh=0.007, n=14, F1 Score: 77.48%\n",
      "Thresh=0.009, n=13, F1 Score: 72.32%\n",
      "Thresh=0.009, n=12, F1 Score: 72.44%\n",
      "Thresh=0.010, n=11, F1 Score: 72.27%\n",
      "Thresh=0.011, n=10, F1 Score: 72.18%\n",
      "Thresh=0.011, n=9, F1 Score: 71.77%\n",
      "Thresh=0.015, n=8, F1 Score: 72.23%\n",
      "Thresh=0.021, n=7, F1 Score: 74.09%\n",
      "Thresh=0.039, n=6, F1 Score: 74.09%\n",
      "Thresh=0.044, n=5, F1 Score: 70.77%\n",
      "Thresh=0.054, n=4, F1 Score: 70.77%\n",
      "Thresh=0.107, n=3, F1 Score: 70.32%\n",
      "Thresh=0.223, n=2, F1 Score: 65.68%\n",
      "Thresh=0.303, n=1, F1 Score: 34.21%\n",
      "CPU times: user 3min 39s, sys: 1.49 s, total: 3min 40s\n",
      "Wall time: 3min 41s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "from sklearn.metrics import f1_score\n",
    "\n",
    "# we'll use the data above to re-run the model but go through various thresholds \n",
    "# to see if we improve performance by getting rid of various features. \n",
    "\n",
    "thresholds = sort(model_two.feature_importances_)\n",
    "\n",
    "for i in thresholds:\n",
    "    selection = SelectFromModel(model_two, threshold = i, prefit = True)\n",
    "    select_X_train = selection.transform(X_train)\n",
    "    \n",
    "    selection_model = xgb.XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "              colsample_bynode=1, colsample_bytree=0.8,\n",
    "              enable_categorical=False, gamma=1.5, gpu_id=0,\n",
    "              importance_type=None, interaction_constraints='',\n",
    "              learning_rate=0.02,objective='reg:logistic',  max_delta_step=0, max_depth=5,\n",
    "              min_child_weight=1, monotone_constraints='()',\n",
    "              n_estimators=600, n_jobs=1, nthread=1, num_parallel_tree=1,\n",
    "              predictor='auto', random_state=0, reg_alpha=0, reg_lambda=1,\n",
    "              scale_pos_weight=0.237, eval_metric='auc', subsample=0.6,\n",
    "              tree_method='gpu_hist', validate_parameters=1, verbosity=None)\n",
    "    \n",
    "    selection_model.fit(select_X_train, y_train)\n",
    "    \n",
    "    \n",
    "    # model evaluation \n",
    "    select_X_test = selection.transform(X_test)\n",
    "    threshold_predictions = selection_model.predict(select_X_test)\n",
    "    f1 = f1_score(y_test, threshold_predictions)\n",
    "    print(\"Thresh=%.3f, n=%d, F1 Score: %.2f%%\" % (i, select_X_train.shape[1], 100*f1))\n",
    "    \n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b9622170",
   "metadata": {},
   "source": [
    "It appears that once you get past 13 features, the additional features don't appear to add much value to the model, because while there are some fluctuations, the F1-score stays around 77. Given the feature importances, it appears that interventions that would reduce defaults would be fairly simple: make the loans more affordable to mitigate the problems the customers are having with keeping up with their expenses. \n",
    "\n",
    "Consider the following:\n",
    "\n",
    "* A group of 1,000 customers in danger of default, who represent a potential loss of 7.5M based on the average loss of 7.5k per bad loan \n",
    "* Let's say all 1,000 customers are offered interventions of ~1.5k/each, as far as reduced payments, interest rates,etc and half take the offers and use the entire incentive and of those 500 customers, only 100 avoid default. \n",
    "    * Total intervention cost: 750k\n",
    "    * Avoided losses from the 100 who pay off their loans: 750k\n",
    "    * Given that you would only need to reduced defaults by 10% for interventions to be revenue neutral, plus the potential gains from the other customers who take advantage potentially making more payments, there is strong potential for significant avoided losses. \n",
    "    * A good approach could be just structuring interventions so that you only need a 10% success rate to come out ahead, and/or structuring them so that customers have to take make several payments to receive the entire benefit. \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fa449a9f",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2 µs, sys: 0 ns, total: 2 µs\n",
      "Wall time: 4.29 µs\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# for reference, we'll run the model via logistic regression \n",
    "# commented out the code since we won't use it, have included a screenshot of the results\n",
    "# as you can see the XG Boost are slightly better, but can be generated in about 7 seconds (with GPU)\n",
    "# vs. 11 1/2 minutes for the model below. \n",
    "\n",
    "#from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "# l1 \"Lasso Regression\" logistic model, which uses an 'l1 penalty' \n",
    "# the penalty effectively scales with model complexity to force the model to generalize and avoid overfitting \n",
    "# l1 penalty is called the \"absolute value of magnitude\"\n",
    "# An overview of the penalties can be found here: https://towardsdatascience.com/l1-and-l2-regularization-methods-ce25e7fc831c \n",
    "# The balanced hyperparameter is used to handle the fact that the dataset is unbalanced \n",
    "\n",
    "# create the model \n",
    "#lr = LogisticRegression(penalty='l1', solver='liblinear', class_weight='balanced', random_state=123)\n",
    "\n",
    "# fit the model \n",
    "#lr.fit(X_train, y_train)\n",
    "\n",
    "# generate predictions and a confusion matrix \n",
    "#y_pred = lr.predict(X_test)\n",
    "#cm = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "# display the confusion matrix and print the classification report \n",
    "\n",
    "#print(cm)\n",
    "#print(classification_report(y_test, y_pred))\n",
    "\n",
    "# print default rate \n",
    "\n",
    "#fp = cm[0, 1]\n",
    "#tp = cm[1, 1]\n",
    "\n",
    "#default_rate = fp / (fp + tp)\n",
    "\n"
   ]
  },
  {
   "attachments": {
    "l1_regressor-2.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "d57aa401",
   "metadata": {},
   "source": [
    "XG Boost only delivers slightly better results, however, since XG Boost can be accelerated via GPU the results are delivered in seconds vs over 15 minutes for an Sklearn L1 Logistical Model. Note: this model delivered better results than decision trees, or an L2 regressor. \n",
    "\n",
    "![l1_regressor-2.png](attachment:l1_regressor-2.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9c168027",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2 µs, sys: 0 ns, total: 2 µs\n",
      "Wall time: 4.77 µs\n"
     ]
    }
   ],
   "source": [
    "%%time \n",
    "\n",
    "# commented this out as it takes a rather long time, \n",
    "#screenshot below shows the results from one of the times the model was run\n",
    "\n",
    "#cv = StratifiedKFold(n_splits=10, shuffle=True, random_state=1)\n",
    "\n",
    "# evaluate model\n",
    "#lr_results = cross_val_score(lr, X_scale, y, scoring='f1', cv=cv)\n",
    "\n",
    "#print(\"F1 Score: %.2f%% (%.2f%%)\" % (results.mean()*100, results.std()*100))\n",
    "\n"
   ]
  },
  {
   "attachments": {
    "lr_cross_validation_results.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "ea70f587",
   "metadata": {},
   "source": [
    "Similarly to the L1 regressor vs. XG Boost, running cross validation with Sklearn took several hours due to the unavailability of GPU acceleration, while the same training and validation method took a little over a minute with GPU acceleration using XG Boost. I.e. beyond the slightly better results, XG Boost is just a vastly superior option as it enables a data scientist to rapidly try different approaches, data and hyperparameters without needing to wait hours to get the results. \n",
    "\n",
    "\n",
    "![lr_cross_validation_results.png](attachment:lr_cross_validation_results.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94363375",
   "metadata": {},
   "source": [
    "#### Final Take Aways \n",
    "* The model delivers significant value in terms of increasing capital efficiency \n",
    "* Only the top 13 features provide value in terms of predicting defaults, which appear to largely be a function of interest rate, post loan debt to income ratio, credit card balances/usage and the payment amount\n",
    "* Potential interventions to reduce defaults should be focused on reducing payment amounts and interest rates; even a small reduction in defaults could easily cover the cost of the interventions.\n",
    "* The model needs more data, both in terms of volume and in terms of having more data points with strong predictive value, as there were over 40 data points that added almost nothing to the model. "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "GeneralDS",
   "language": "python",
   "name": "generalds"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}