Email_Opt_In_Prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "78c1390f",
   "metadata": {},
   "source": [
    "# TABLE OF CONTENTS\n",
    "\n",
    "####  INITIAL DATA PREPARATION\n",
    "* [Initial Data Imports, Variable Explanations & Merging](#initial_data_import)\n",
    "* [Initial Data Exploration & Visualisations](#ide)\n",
    "* [Data: Dummy Variable Creation](#data_pre)\n",
    "* [Feature Engineering & Data Cleaning](#feature_eng)\n",
    "\n",
    "#### Classification Models\n",
    "* [Un-tuned classification models comparisons & cross validation testing](#models)\n",
    "* [Tuned classification models comparisons](#tuned_models)\n",
    "    * [Additional Voting Classifier Approaches](#add_vc)\n",
    "* [Tuned Classification Models on Test Data (out-of-sample) Prediction Accuracy](#oos_models)\n",
    "* [Final Model Prediction Results Summary and Discussion](#model_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3d29aabf",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "attachments": {
    "image-4.png": {
     "image/png": "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"
    },
    "image-5.png": {
     "image/png": "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"
    },
    "image-6.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOIAAACHCAYAAAALKYU+AAAVvUlEQVR4Ae1dva7cuA6eNwxwgLzOIgjOg2SLk/6+QbY5KVNsu0WALbZZYMuUviApSiRFyfaMf8cMEFi2JJL6yI+SZXvObRiG4d9//x3+/PPP4cePH/E/MIgY2CEGbkBEIOF///03/Pr1K/4HBhEDO8QAEhFmwiBhJKGIgf1iIIi4Q/aLgN8v4I+KfRAxiBiroQPEQBDxAE44apYOu7abuYOIQcSYEQ8QA0HEAzghZp7tZp6jYh1EDCLGjHiAGAgighP++DTcXr4MPw/gkH7G/jZ8ut2GG/7/MHz5a8eZxMUs2XcKLHfEzomzfYn415fhw+3T8M0xrB+QC4PoBtXCOhYdIwR8g4hbjcXV8wARjxILi/ppegwFEQF4N6img7hp0sBAOSoRH8AsiDjgu4XtYPo5fHnh5RAceQaj7PfpDwm+d03WczllzrzMYvkly//8/cNw++1buX9RjgKboK2Uw3YlHUCuLL/I5XGi/Fx/m7E0JTxg3N9+q+1G+V3dLTzBbhiPthV0fPj9Z8GhRUTEh+2RR4lLTzf7pn3sYabqpN/EDKPa3G4DxY70obRb48B+e9bjyIyYHNcAFgNR1imytB2awey0R6c1ZXNAsbPoPAcsyuW6NOPlBNI4n3xfw7oLQRQOI7qrcYlAfYiILKczu/d1j/gLk4sgdUNPSwdelz5ge/nYiYUcL9z2CY99IlrwLQAGPAA7k8G29c5Nfwl45VDVlsggZ2PZvrZDtpflFHyNoJL2lHIiokwSon9f968B7WwG5AMzIuMrbCk20zj7untEnI4Z6pDYoF0060l/Wdt+Kf/2bHnOui4REdTuTCEdBGUxC3Fg9I4d8CuHqrZSrwiyFABlySiXOrwUqvvOu0d0+osx9nWTrbKNTlzrEhGCv627F+DOmBuEr/wG2KDvRmJD+bdny3PWdYk4JUAz8ABkl7QOgB3ws1wOctW2DgzZHoJNB7jUXfedMs6SwZ3+bGMK9LZuaQeUaaYo7dcnYhmH1W1tk+fOmOcQ0bn3LXYkPcq/Uvc1yn0iVoHigJII+O33D+nm22kjAlU7gILBW7IgsTKxqV3ZKKoDQxIRiWU2PaRenBV4+YQBcN9mjZSZyxCgHd25HWJC49BE5Jmbl7FeUqkJm+XieEZmH1d3329TMVN+EH7H/tmfnq52LOSxCXnPdm2EiAAYk4CXeeKGPQGDIDfvezzQxTUMXJYtA4iCND+8/kM+cxwhItil5IJ8abcc06fhW0om0x7o17qroGjqlmNKY+aEwEEm+/72DZeSmaiyLu/4SswIVyRDrudxT9DNNrjHHmaObNBvxkZxwr4uCSfjp8ZXjyu3c+0TMXXC+glEPPcAn9l5Mbbnic0g4gmzZxDweQjIvlyHiHzflZdHZTkCS8281DocCeTyS9uMS2Sz1GIQz37US1k77mstEffy5TpEPBzBni+D7hUwoXedWAoiRtIwr8+tE2hB4D6uQcQgYhDxADEQRDyAE2K26M8WV8AniBhEjBnxADEQRDyAE66Q8WOM/VkfiRg/ud8HKYIo8Fk7BpCI8Udo4o/vxB8g2jcGkIjwh2jiXyAQCOyHQBBxP+xDcyCQEQgiZiiiEAjsh0AQcT/sQ3MgkBEIImYoohAI7IdAEHE/7ENzIJARCCJmKKIQCOyHQBBxP+xDcyCQEQgiZiiiEAjsh0AQcT/sQ3MgkBGYTsS/34YX/OmLl+Ht79w/Fd6H19ttePn6j604zvn31+H28W2YZiGNh35BzhvvcYblW3ICf/iGX/bqRCL+M7x97BGt43gk8OvwvjfEs4jIxsK4goiMxjBQHFCCug23z/O9+v5Z/CaO6K+u327D6/eiFUq2/ib9kicJX7aWdMyziUQcC0giogUPhxxE3MHzHX88YA2SIZOHdExfBSUS5/4dQzBmdAIE3W58gRhoP3m109G7Y9VCRPRGQI7K2TP/ohsDTI6R4P7z9UVkWej/OrzLbGfBhlmukltsQXm5/jZjacoywAa2l6/RUWdo28YbO7cp4y4yuC7p6I6rJ1vb2DojXIzOVmO+7pAD5VifcHt7hDFNISH2q3EPIjaAsTg3z9GB3tK0BCT3rYkISw3uSwGYiWsDA4OX2w7D4J1PDRo2aKgDAqpsAFJgF91IMA66lEiy3WJ5x7NJ3V6QxIyjbtuZKfI4dIHsFTp0tX8Gdkj80C7pH78bXwW7X76+4V4CJ8+CCbdKR6srLU2b7QFjaZsRd4bTaTNik0wThtjsO4WIOljImbTdAsHEgUxWSHmynGx0nDtuvUdEkxBQiNRn66mu2ErnanYQtvXHNSZ7fER3t2AbU2KhBOnh42lIY85JlROl9C+3ofs8SzpMQGJ1U/BMS1NRV5K3Z8sxr40QkRyvMuHccaxEROsYnWXJqcqZHEiz7PcCzbsm9aWAUjNiHXDKNmHTlHFlEiO2UrYQtHQR8MNgLzM/3ptJcjV1Sny4kXdN1ymycRUeKS5b9XaForoe9GSEiMnqJpkmjKrZt3YEAsgB7CwL5Ywoy7UVtWxcqs5evnikoyDQRNLXLJl0wDi2iQH0x1XvHmrZQtDSRY/0M5JbPa4+DjoW6sHU8kSbZsyJNgcrTiOiQ4rp49BBWvqRI3IgccadSES6B2zPBkgGloWOWW6zBmULUmPQ8PloEPQDsDuuUdkF3V6JZow2dn5fsrusjoz/UifExnn84N6zN2fTVswkJRgrLfuTnex7HszdMcAC1j1uQES+H+BnPAJABgeWPB/fhvdq11S0TTfsmbiAC5M33x+IZRMmD9aZdl+ZLGOYVnJBTm0LL4ftPQkHY6mXz2BHiDgyrr7ssYFR/X1EhL4pyBPeyhdJNdunVwxSr/BJ6mPlAm66PxGz4Cn9nDbPcgxIrLMCYbvuK1vsWd6GiHuOcGvd3qyF1zSR7zJrTdl3GXSuTpgkpibjjYc2kYgTsvjGhh9WHc6mOuvSDKSv3WX/mrLvMugknXiFc1ASAooTiSi3iBfI7Lv6zy5zeKmUjvbe4g5beXnWWkrdITJ3WVN2VhKFzRGYTsTNTQuFgcB1EAgiXsfXMdIDIxBEPLBzwrTrIBBEvI6vY6QHRmAhItIGiPdc6TBjh52zqbtmzu7k8uM4AWbLDzokNhBYn4jes6+GMatevgwR9a7w9OSoH9aXHd+yS253bPVD9+I9bqfq+RECP3hfYHe6aDx/aVEiKuAZmyAiI2GORBgXM9Ny+imRKZMPsbdvqUyXVr2WJrui7ELSXIUJ73V4/djTa+zMna9bWIiIHoA6M3sZFh90y8yoSAv9d/ow2CxN6YF8eXUKziHYOfPD2AqhIMggQOX45z/MJ51OoHtQ8zVjN1xGGyXG3HbCEfpmUlftYXzWPhrz63ciWsGk6ox2tWXX7Z/9yopETNApcmk4x4kID9k5iNnJUq4IBBuE3vkd94hECLaBdNO1EqR4nmVTEJZ3U+/L/qRDjE9D555hH0E6tnPyvbGU2vEbNgN885ipYyH9GBE9Ekvl1ysfnIg6EGWGhiDTGVU6X5aTU53Aabo7kfgNXkLPiaC0tgGvv8urdVfti6hFS1kP2p++NhkjVMMCiXVpQmPj1Y2a8ZSeGgOQgfbFPWKBU5ROS0TMvuxUcaTgcAJhNhHbr7zlgGcgR4Kwas/9Fj7mQJcz1Zxxsz04Hp0EuaocCWNKhhZve156cQltlXZyxUWPpyainhGlB51AmBOQOKPAcpSWw1ZPRayDENHbXKlslTA1ymWJ2WiQLmfZOH7zzi4nxxbZFGZ9PVeoXZ+IKZjVMkY6MjuKgr4sBev7CLVcQrK0s7YKJg6UrGvEtZmI5WV3aX8OQBajgqpOAlV77tc5Yp9qM6TTAasShvk+kc6l7dCMVxP2OorAsbRxLRb4sqm+xqD0o5LyD1fO9RP3e4LjBkS0H/BKJ5PD6J7jZXj7Dr8mzhsjI0QE8JEwMhNzX6hMQYmZ+Z4Pg4Us1pMCvCLWYYhoxy13c0u0tomY/JGJXPqMf7xbt5VEp8QifNXVIbCXYp+4vA0RnxjAGNqyCGCSmLpyWVb1rtKCiLvCH8ozArzquCAJAYMLElEuWcVSiTcX3CVTDpcoBAKrIHBBIq6CYwgNBB5CIIj4EHzRORBYBoEg4jI4hpRA4CEEgogPwRedA4FlEFiIiLQBYt9AWcbEhaTArtziO3J642fe+E+A2ULQh5hxBNYnonrYPW7Qai1WIKL7dsjkAaxFxEeSQ3nrxn2xm8fWeANGP7Q3D+X58UTsTjOK6rgoEeWbFFnL0xJx/DWujIFbIMK4mLntp1wkm/LMnAgzVUf1xhCSxxAq/ez+62ezwjBtUVZzBWLsnDK0J2+zEBE9lHRm5gxbvtNLn8XI53aKtNB/pw+Dsx1iDFVQjRDJzACZHB5UzjWaXeTrgE4je8mQAarnzNrQVtmZcSiKMsFAV8bEIxbg07a/0lVUXLK0IhETno4zGekqA6u2TALOyCbwsa1wtA1C7zwHDlvQOKJseNjP8oXuXGdfBmA7+R1Y7tvQMXL5HiJaPEnGjL+ChZjx+6lErvy3GMFeiblDRD3zUn99jQfdJym3utLx4ETUwSyzKASZyt68ZPoO7nOCQAXOiIsT2WQQSd3UW5BTifNmB9VgtZNMxEQonLFwLCJJjGqnccEKRuNLs2u+ZvBE3SLRcRKQGPI1XB3JldCoTc/f4LREBGKU5W4pk+OXIOJY8LaI2Lq+fjDlQBeEwC9U5HnPDCQwJ79ESO5riFfLJczZJy9f39Nv9/gKLXH9Vte5emoi5uxc+WtPIjq6K/tWuoBE0gkEA37S7EN2a0xLUmklPiCenPXKyKCvtqXU8TK3U68aP//J+kQ844fBk5ZzJUhtmNDM9FiQkQyenayG1nmaxTLxfBuZVJpAaUbLfUfude0MqUzy9comaIPUBZXplqBsAskez13egIjsUF4+yuBKzsdnSwf6MPhBIkLIEJF4zPX91lhY3UdEkJrImJ7XabKRVp+Idd+yWeVYa4mIszGPV/qY+lo81CZQFs/x8FgSy+JOVNiGiCcCJEzdFwFMEnxfuq8pm2oPIm4KdyhrIsAz6gVJCJhckIh66ca7fPlo71uakRMVgcByCFyQiMuBF5ICgaUQCCIuhWTICQQeQCCI+AB40TUQWAqBIOJSSIacQOABBBYiIm2A6LcyHrBqja72udcaOlAmPQvznt9plSfATBscZysisD4RJz0cX3GELPoyRNS7wnOTIz/s511kN6GMvAHDMry+VFc/8Gc3XfW4KBE94PWfLNsR5oMS0cXsbphoNs7kS4SZqgPffpGPb/DZnn3LJc349sNgthlx9v5iMCWI16/wZxWCiAwXHxciIouTR52ZOcPK16Yqx6vZE/rv82EwZW0RgCmgc4DzB7f8sw/pSPW8NJXjF7IkRJ3yXa+4OcTBsUhydXRCWzlGL4miXfDQ3U1siWzOXwx+/5zIh1gGEa0bViRiUqXIpdWPExHeXeQgZidLucKhNgi988lvbRCZ6H1IM8uAeiVbtoXKdC6y/hwyMEL3ENHiSTK2+DCYrC7jJAzcmTiIyC5Wx4MTURAtzUKcsSHIuEwjks6X5TReN4MrLMwJEP9lePsqfxKCmpSAo3MM+Ezyhu6cUIyaBU8zETFR3PsXg8tMrvHtfxisZ08HAx5nEJGRUMfTEhHJYJaGsPylLOwEwmwi8synkwGgRzMNz9SkqwSto7uzKlDeePCE7DIz4JxxI4F5vImQnGCsHHVux2zPxcCCiAKMUjw1EUvwlwFRyQkEFTi2vXPO5MHgZNKldniNP/kxgf/oz3Q4pky+5NiaZ8lRIYSZxrTcDrQSHya//8EGjMBDlpnIrD+IyEio4/pEPOOHwYlMHJR6KeqQXEFq670AVx3cE5rdeHZymzgX0yyWN2cKkWRjJpW+hyM71XeCSOyGDVBnSZaVWAxyhf4BKnEZi0hSm9hso+c834CIvMTjjCkdm5yPGfQoHwYnm2SQcYBwgNsZEezP7eWYaMxM6DkhdB8RQUMiY5qVNNnIAp+IdV+5w13ZPpOIedksZ0uxoUXyGTuzAqmUP9+FbYj4VLhRsOgAp2v3EO6poFlgMJgkclJbQOBJRAQRZzvKWe6lGVOTc7bga3fgVcYFSQiOvyAR9dKtvGiQls68/OzRgoNGLLOChD3Aom4MgQsScQySqA8EtkcgiLg95qExEKgQCCJWkMSFQGB7BIKI22MeGgOBCoGFiEgbIIfevu8+96pw2eDCCTDbAAVWkZ8zxq4pQ3LPsRNUuLV/gAe0lyGi3hWemxz5YT/uJlek4AfujR1mtZssX9ygmFKy83vBOt6gzVybtYRzni06I7pb+EHERmQQYVzMGj3GL5sXC2Y+38RZSZAPiSMe5+hzk3xRVyEfzXAlAeO5kKU/JRMjAzLLdqLqmYsLEdGDSGfm8rzOOEuCrkgL/ff5MBhHo7K7zNIegZxrzf4eVvU1CuSCVd3CuYI6S/BDC00ep0++5IxB+sMQDbpJ4tZ6tDyoVzOdlJ1tSK9DypiQdU9cXpGICbUW4OxICbpqS47c58Ng0N0mQRV0ym5+t7bdf0o83UNEO+uQDPkebE+zN2ZxDUguZkua0WCJCsS3r/2VJWwmX0pMtAJI9dL3bBq0865z/ZMeD05EHcwyq0KQZSejc2QwyHLynA2krkMpCWj5ooMhnraFdDf7CjFLFzMRU9AjcYytPZ02weA5JyTGD+UxAZmoBW/qQ7NfticrJVxhddTEZ4a9WewTFE5LRHZ4WfLSBoLMuOr+iwNpqtNywHHQyY4l8OinMWTCoGBTumXXFcsY+OpLkDQ7y5msq78QBXB9/Q7naanL5OZzkJNJQ3hYgoGPMuGwP+OU9LTsYuwvNDOemojZyVVwSaKkyrlEFDIpwPW9V872EDQqoBzdQtaqRQz2hp33KJZjQ3IwkZIwgSkmRkUciQOVtb8aCUvIvMfks/ZZn4in/DDYuNMJcJwNPr4N719fqj9d7RHXSBw9JRkm8Ed7pZkmE8IPdl5NdGftinhEpnKfaMiFGPFPlfB9MieF1DfbxfXO+ECObDc65udosAERGXRaOuqPTZOD8CuGo3wYbO0Fu52A4R1JuVQTMUFE4jGLJZpo0yveR0SQmMiYvgzxyNYioraZSSStlP5yxpTISLcLtr+2q4UpbgIFESXoUQ4EdkAgZsQdQA+VgYBBQG3wmLpnPt1maXooBO0SqSwfcUl1wWXREdxTlsV2SXsE69a34YJEXB/U0BAIzEUgiDgXsWgfCKyAQBBxBVBDZCAwF4Eg4lzEon0gsAICQcQVQA2RgcBcBIKIcxGL9oHACggEEVcANUQGAnMRCCLORSzaBwIrIBBEXAHUEBkIzEUgiDgXsWgfCKyAQBBxBVBDZCAwF4Eg4lzEon0gsAICQcQVQA2RgcBcBIKIcxGL9oHACgj8H8y3v6tTo983AAAAAElFTkSuQmCC"
    }
   },
   "cell_type": "markdown",
   "id": "02fc23b1",
   "metadata": {},
   "source": [
    "## BRIEF PROJECT OUTLINE: \n",
    "\n",
    "### PURPOSE OF CLASSIFICATION MODELS & PROJECT \n",
    "* The overarching purpose of these models (this project) is to determine why some loyalty customers have chosen to opt into email marketing and others have chosen to opt out. The data shows that those customers who opt in have higher average spends, thus it is important to determine whether customers who are already spending more choose to opt in or whether opting in results in a higher average spend. To determine this, I construct various data features which capture demographics of customers and utilise various classification models to determine outcomes based off the input demographics / data features. A desired takeaways is to understand what features underpin each group and inform decisions on how to influence customers to opt into email marketing. \n",
    "\n",
    "#### THE  DATA: \n",
    "* Original data is 52K entries of sales data across 12+ departments and several hundred unique customers with unique individual characteristics: post code, age, gender, shopping frequency and average spend across different departments. Data ranges from 2019 to early 2021.\n",
    "\n",
    "#### PROJECT FLOW:\n",
    "\n",
    "* After initial data exploration and data cleaning, I create a variety of features for the classification models:\n",
    "    * Customer Duration (Time between first and last transactions)\n",
    "    * One-Hot Encode post code data by customer, customer gender, department sales frequencies by customer\n",
    "    * Standardise these variables without mean (Ex. post-code) to preserve the sparse matrix nature of the data\n",
    "        * Also standardise: age of each customer\n",
    "        \n",
    "        \n",
    "* A severe class imbalance problem was present between customers who opted into email marketing (majority class) vs those who didn't (minority class). To remedy this, I upscaled the minority class with replacement to balance the two classes\n",
    "    * Also tried downscaling the majority class in a separate iteration and found the former method to be superior\n",
    "\n",
    "\n",
    "* I first run the following models before tuning and compare performance (accuracy) across the average of (5) cross validations on the training data\n",
    "    * Naive Bayes (baseline measurement)\n",
    "    * Logistic Regression\n",
    "    * Decision Tree Classifier\n",
    "    * Random Forest Classifier\n",
    "    * K-nearest neighbors\n",
    "    * Support Vector Classifier\n",
    "    * XGBoost Classifier\n",
    "    * Soft & Hard Voting Classifiers\n",
    "\n",
    "\n",
    "### UN-TUNED (INITIAL) MODEL RESULTS\n",
    "![image-5.png](attachment:image-5.png)\n",
    "\n",
    "\n",
    "* With the exception of the naive bayes, I tune the above models and re-measure the performance the same way\n",
    "    * I also experiment with different weightings of soft voting classifiers \n",
    "    \n",
    "    \n",
    "###  TUNED CLASSIFICATION MODEL RESULTS     \n",
    "![image-6.png](attachment:image-6.png)\n",
    "\n",
    "    \n",
    "### OUT OF SAMPLE PREDICTION RESULTS    \n",
    "* The final result is the individual tuned classification models perform best; consequently, I perform predictions on the test data with the trained model to measure out-of-sample predictive strength\n",
    "\n",
    "![image-4.png](attachment:image-4.png)\n",
    "\n",
    "\n",
    "\n",
    "###### FOR FULL RESULTS REPORTING AND VISUALS, FOLLOW THE ABOVE HYPERLINKS FOR THE RELATIVE SECTION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "19e46a16",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "568dfbcf",
   "metadata": {},
   "source": [
    "## PACKAGES\n",
    "* NOTE: some additional packages are imported in the \"classification\" models section, concurrently with the introduction of each model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3358afe2",
   "metadata": {},
   "outputs": [],
   "source": [
    "## import packages ## \n",
    "import warnings\n",
    "import sys\n",
    "if not sys.warnoptions:\n",
    "    warnings.filterwarnings(\"ignore\")  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "33944149",
   "metadata": {},
   "outputs": [],
   "source": [
    "# core packages\n",
    "import numpy as np \n",
    "import pandas as pd\n",
    "\n",
    "## graphing packages ## \n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "\n",
    "## date working / printing packages\n",
    "import datetime\n",
    "from datetime import date, timedelta, datetime  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "6edf7357",
   "metadata": {},
   "outputs": [],
   "source": [
    "# model packages \n",
    "from sklearn.metrics import accuracy_score \n",
    "from sklearn.preprocessing import MinMaxScaler, StandardScaler, LabelEncoder\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "918079fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "## optional packages / display preferences ##\n",
    "from IPython.display import display\n",
    "pd.options.display.float_format = \"{:,.2f}\".format\n",
    "np.set_printoptions(precision=2)\n",
    "pd.set_option('display.max_columns', 50)\n",
    "plt.style.use('seaborn')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5035d904",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bf75a4ee",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "285f9609",
   "metadata": {},
   "source": [
    "<a id='initial_data_import'></a>\n",
    "\n",
    "## IMPORT FIRST DATA FILE: SALES DATA\n",
    "* E.G. data contains sales by department, by customer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "11c5768e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import xlsx file\n",
    "weekly_customer_sales_bydep = pd.read_excel(r'weekly_customer_sales_department_data_mw.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "62ca9d34",
   "metadata": {},
   "outputs": [],
   "source": [
    "# [optional] temporarily need REGEX; then can del re\n",
    "import regex as re\n",
    "\n",
    "# CLEAN DF COLUMN NAMES OF SYMBOLS / SPACES: \n",
    "# make all characters lowercase, remove spaces etc.\n",
    "for i, val in enumerate(weekly_customer_sales_bydep.columns.values):\n",
    "    weekly_customer_sales_bydep.columns.values[i] = re.sub('[^A-Za-z0-9]+', '', val.lower())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "85271529",
   "metadata": {},
   "outputs": [],
   "source": [
    "# === NOTE === # \n",
    "# to protect the anonymity of the data, the actual store_name has been replaced with a code word\n",
    "# this store name has been hidden below, replaced by the string \"####\"\n",
    "weekly_customer_sales_bydep['store'] = weekly_customer_sales_bydep['store'].replace(\"####\", \"MW\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "ca86dfe8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# for privacy concerns, we keep identify a unique customer by their card number and hide their full-name\n",
    "weekly_customer_sales_bydep.drop(['fullname'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "febad655",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 51524 entries, 0 to 51523\n",
      "Data columns (total 9 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   store            51524 non-null  object \n",
      " 1   weekendingname   51524 non-null  object \n",
      " 2   sales            51524 non-null  float64\n",
      " 3   baskets          51524 non-null  int64  \n",
      " 4   sumofcardnumber  48741 non-null  float64\n",
      " 5   sumofage         14926 non-null  float64\n",
      " 6   gender           48741 non-null  object \n",
      " 7   city             16414 non-null  object \n",
      " 8   department       51524 non-null  object \n",
      "dtypes: float64(3), int64(1), object(5)\n",
      "memory usage: 3.5+ MB\n"
     ]
    }
   ],
   "source": [
    "weekly_customer_sales_bydep.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2efae73a",
   "metadata": {},
   "source": [
    "* NOTE: a lot of missing data around age and city (post-code)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "d93a5acc",
   "metadata": {},
   "outputs": [],
   "source": [
    "# rename some of those long/verbose column names for simplicity \n",
    "weekly_customer_sales_bydep.rename(columns = {'weekendingname':'week_ending',\n",
    "                                             'sumofcardnumber':'card_num',\n",
    "                                             'sumofage':'age'}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "4d700e8e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "store              0\n",
       "week_ending        0\n",
       "sales              0\n",
       "baskets            0\n",
       "card_num        2783\n",
       "age            36598\n",
       "gender          2783\n",
       "city           35110\n",
       "department         0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# visualise Null values for the DF\n",
    "weekly_customer_sales_bydep.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1dd1666f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "e3705222",
   "metadata": {},
   "source": [
    "#### Remove sales entries that don't have a corressponding customer (card_num) attached\n",
    "* Other category NaN - E.G. Gender, will be handled in the feature creation section just before the classification models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "d178ed20",
   "metadata": {},
   "outputs": [],
   "source": [
    "# we see at the begenning of each \"week_ending\", there is aggregate department-level store sales information we don't need\n",
    "# there is no associated card_num with this information obviously, so we can drop it\n",
    "weekly_customer_sales_bydep.dropna(subset=['card_num'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "6d853b4b",
   "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>store</th>\n",
       "      <th>week_ending</th>\n",
       "      <th>sales</th>\n",
       "      <th>baskets</th>\n",
       "      <th>card_num</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>city</th>\n",
       "      <th>department</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>MW</td>\n",
       "      <td>Jan 06, 2019</td>\n",
       "      <td>8.58</td>\n",
       "      <td>3</td>\n",
       "      <td>1.112257e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>BREAD</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>MW</td>\n",
       "      <td>Jan 06, 2019</td>\n",
       "      <td>34.75</td>\n",
       "      <td>1</td>\n",
       "      <td>1.112257e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CIGS &amp; TOBACCO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>MW</td>\n",
       "      <td>Jan 06, 2019</td>\n",
       "      <td>4.99</td>\n",
       "      <td>1</td>\n",
       "      <td>1.112257e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CONFECTIONERY</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>MW</td>\n",
       "      <td>Jan 06, 2019</td>\n",
       "      <td>11.44</td>\n",
       "      <td>1</td>\n",
       "      <td>1.112257e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>DAIRY</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>MW</td>\n",
       "      <td>Jan 06, 2019</td>\n",
       "      <td>26.62</td>\n",
       "      <td>4</td>\n",
       "      <td>1.112257e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>FRUIT &amp; VEG</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   store   week_ending  sales  baskets      card_num  age   gender city  \\\n",
       "23    MW  Jan 06, 2019   8.58        3  1.112257e+17  NaN  Unknown  NaN   \n",
       "24    MW  Jan 06, 2019  34.75        1  1.112257e+17  NaN  Unknown  NaN   \n",
       "25    MW  Jan 06, 2019   4.99        1  1.112257e+17  NaN  Unknown  NaN   \n",
       "26    MW  Jan 06, 2019  11.44        1  1.112257e+17  NaN  Unknown  NaN   \n",
       "27    MW  Jan 06, 2019  26.62        4  1.112257e+17  NaN  Unknown  NaN   \n",
       "\n",
       "        department  \n",
       "23           BREAD  \n",
       "24  CIGS & TOBACCO  \n",
       "25   CONFECTIONERY  \n",
       "26           DAIRY  \n",
       "27     FRUIT & VEG  "
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "weekly_customer_sales_bydep.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a2b4146e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "fbe9972e",
   "metadata": {},
   "source": [
    "### [OPTIONAL] \n",
    "* Pickle the DF thus far"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "5a849da8",
   "metadata": {},
   "outputs": [],
   "source": [
    "weekly_customer_sales_bydep.to_pickle(\"./weekly_customer_sales_bydep.pkl\")\n",
    "#weekly_customer_sales_bydep = pd.read_pickle(\"./weekly_customer_sales_bydep.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9aac5952",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a3aa2a8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "54cc15fe",
   "metadata": {},
   "source": [
    "## IMPORT (2ND DATA FILE): CUSTOMER INFORMATION FILE\n",
    "* Contains relevant information like: email marketing allowed, post code etc. \n",
    "* This data will be merged onto the above file \"weekly_customer_sales_bydep\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "75788d8d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import supplementary file with customer information\n",
    "customer_emailopt_in_file = pd.read_excel(r'mw_customer_information_emailoptin.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "1935efc8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# once again, drop the full_name for privacy reasons \n",
    "customer_emailopt_in_file.drop(['Full Name'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "cc7981be",
   "metadata": {},
   "outputs": [],
   "source": [
    "# to facilitate the merge, we will rename the relevant columns \n",
    "customer_emailopt_in_file.rename(columns={'IsHomeEmailMarketingAllowed':'email_status',\n",
    "                                         'Sum of Card Number':'card_num',\n",
    "                                         'PostCode':'post_code'}, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1122093",
   "metadata": {},
   "source": [
    "## E.G. Of the two DFs data for one individual customer (card_num)\n",
    "* Visualise the data available in each DF for a singular random customer with card_num = '59626613925472'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "0acaef18",
   "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>Store</th>\n",
       "      <th>Gender</th>\n",
       "      <th>City</th>\n",
       "      <th>email_status</th>\n",
       "      <th>card_num</th>\n",
       "      <th>Sum of CardCustomerAccountKey</th>\n",
       "      <th>post_code</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Mermaid Waters</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>True</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>4070132</td>\n",
       "      <td>4220.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Store  Gender   City  email_status        card_num  \\\n",
       "0  Mermaid Waters  Female  Miami          True  59626613925472   \n",
       "\n",
       "   Sum of CardCustomerAccountKey  post_code  \n",
       "0                        4070132     4220.0  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_emailopt_in_file.loc[customer_emailopt_in_file.card_num==59626613925472]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "9fee0a98",
   "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>store</th>\n",
       "      <th>week_ending</th>\n",
       "      <th>sales</th>\n",
       "      <th>baskets</th>\n",
       "      <th>card_num</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>city</th>\n",
       "      <th>department</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6017</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.99</td>\n",
       "      <td>1</td>\n",
       "      <td>5.962661e+13</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>BREAD</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6018</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>5.00</td>\n",
       "      <td>1</td>\n",
       "      <td>5.962661e+13</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>CONFECTIONERY</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6019</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1</td>\n",
       "      <td>5.962661e+13</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>DAIRY</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6020</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>15.00</td>\n",
       "      <td>1</td>\n",
       "      <td>5.962661e+13</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SEASONAL LINES</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6021</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1</td>\n",
       "      <td>5.962661e+13</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SNACKFOODS</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     store   week_ending  sales  baskets      card_num    age  gender   city  \\\n",
       "6017    MW  Apr 21, 2019   1.99        1  5.962661e+13  792.0  Female  Miami   \n",
       "6018    MW  Apr 21, 2019   5.00        1  5.962661e+13  792.0  Female  Miami   \n",
       "6019    MW  Apr 21, 2019   1.00        1  5.962661e+13  792.0  Female  Miami   \n",
       "6020    MW  Apr 21, 2019  15.00        1  5.962661e+13  792.0  Female  Miami   \n",
       "6021    MW  Apr 21, 2019   3.00        1  5.962661e+13  792.0  Female  Miami   \n",
       "\n",
       "          department  \n",
       "6017           BREAD  \n",
       "6018   CONFECTIONERY  \n",
       "6019           DAIRY  \n",
       "6020  SEASONAL LINES  \n",
       "6021      SNACKFOODS  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "weekly_customer_sales_bydep.loc[weekly_customer_sales_bydep.card_num==59626613925472].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4d5fd51d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "eeaf1f4b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# filter for the information we need in the merge (don't need gender / city again etc.)\n",
    "customer_emailopt_in_file = customer_emailopt_in_file[['card_num','email_status','post_code']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "17415fb6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0fd7b0b7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "b756a84f",
   "metadata": {},
   "source": [
    "## MERGE THE TWO DFs \n",
    "* merge sales data with customer identifying information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "16cec888",
   "metadata": {},
   "outputs": [],
   "source": [
    "weekly_customer_fulldf = pd.merge_ordered(weekly_customer_sales_bydep, customer_emailopt_in_file, \n",
    "                                          fill_method='ffill', right_by='card_num')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "82b7daa5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 26903 entries, 0 to 26902\n",
      "Data columns (total 11 columns):\n",
      " #   Column        Non-Null Count  Dtype  \n",
      "---  ------        --------------  -----  \n",
      " 0   store         26693 non-null  object \n",
      " 1   week_ending   26693 non-null  object \n",
      " 2   sales         26693 non-null  float64\n",
      " 3   baskets       26693 non-null  float64\n",
      " 4   card_num      26903 non-null  int64  \n",
      " 5   age           13869 non-null  float64\n",
      " 6   gender        26693 non-null  object \n",
      " 7   city          15734 non-null  object \n",
      " 8   department    26693 non-null  object \n",
      " 9   email_status  26903 non-null  bool   \n",
      " 10  post_code     24577 non-null  float64\n",
      "dtypes: bool(1), float64(4), int64(1), object(5)\n",
      "memory usage: 2.1+ MB\n"
     ]
    }
   ],
   "source": [
    "weekly_customer_fulldf.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "efad303e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "7a0d1dc8",
   "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>store</th>\n",
       "      <th>week_ending</th>\n",
       "      <th>sales</th>\n",
       "      <th>baskets</th>\n",
       "      <th>card_num</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>city</th>\n",
       "      <th>department</th>\n",
       "      <th>email_status</th>\n",
       "      <th>post_code</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.99</td>\n",
       "      <td>1.0</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>BREAD</td>\n",
       "      <td>True</td>\n",
       "      <td>4220.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>5.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>CONFECTIONERY</td>\n",
       "      <td>True</td>\n",
       "      <td>4220.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>DAIRY</td>\n",
       "      <td>True</td>\n",
       "      <td>4220.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>15.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SEASONAL LINES</td>\n",
       "      <td>True</td>\n",
       "      <td>4220.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SNACKFOODS</td>\n",
       "      <td>True</td>\n",
       "      <td>4220.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26898</th>\n",
       "      <td>MW</td>\n",
       "      <td>Jan 24, 2021</td>\n",
       "      <td>3.32</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5420602479386</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>SNACKFOODS</td>\n",
       "      <td>True</td>\n",
       "      <td>4218.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26899</th>\n",
       "      <td>MW</td>\n",
       "      <td>Sep 15, 2019</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2710202165626</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>FRUIT &amp; VEG</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26900</th>\n",
       "      <td>MW</td>\n",
       "      <td>Sep 15, 2019</td>\n",
       "      <td>4.59</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2710202165626</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>INSTORE BAKERY</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26901</th>\n",
       "      <td>MW</td>\n",
       "      <td>Sep 15, 2019</td>\n",
       "      <td>4.50</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2710202165626</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>SNACKFOODS</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26902</th>\n",
       "      <td>MW</td>\n",
       "      <td>Sep 15, 2019</td>\n",
       "      <td>2.25</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2710202165626</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>NaN</td>\n",
       "      <td>SOFT DRINKS</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>26903 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      store   week_ending  sales  baskets        card_num    age   gender  \\\n",
       "0        MW  Apr 21, 2019   1.99      1.0  59626613925472  792.0   Female   \n",
       "1        MW  Apr 21, 2019   5.00      1.0  59626613925472  792.0   Female   \n",
       "2        MW  Apr 21, 2019   1.00      1.0  59626613925472  792.0   Female   \n",
       "3        MW  Apr 21, 2019  15.00      1.0  59626613925472  792.0   Female   \n",
       "4        MW  Apr 21, 2019   3.00      1.0  59626613925472  792.0   Female   \n",
       "...     ...           ...    ...      ...             ...    ...      ...   \n",
       "26898    MW  Jan 24, 2021   3.32      1.0   5420602479386    NaN  Unknown   \n",
       "26899    MW  Sep 15, 2019   3.00      1.0   2710202165626    NaN  Unknown   \n",
       "26900    MW  Sep 15, 2019   4.59      1.0   2710202165626    NaN  Unknown   \n",
       "26901    MW  Sep 15, 2019   4.50      1.0   2710202165626    NaN  Unknown   \n",
       "26902    MW  Sep 15, 2019   2.25      1.0   2710202165626    NaN  Unknown   \n",
       "\n",
       "        city      department  email_status  post_code  \n",
       "0      Miami           BREAD          True     4220.0  \n",
       "1      Miami   CONFECTIONERY          True     4220.0  \n",
       "2      Miami           DAIRY          True     4220.0  \n",
       "3      Miami  SEASONAL LINES          True     4220.0  \n",
       "4      Miami      SNACKFOODS          True     4220.0  \n",
       "...      ...             ...           ...        ...  \n",
       "26898    NaN      SNACKFOODS          True     4218.0  \n",
       "26899    NaN     FRUIT & VEG         False        NaN  \n",
       "26900    NaN  INSTORE BAKERY         False        NaN  \n",
       "26901    NaN      SNACKFOODS         False        NaN  \n",
       "26902    NaN     SOFT DRINKS         False        NaN  \n",
       "\n",
       "[26903 rows x 11 columns]"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#.E.G. of the merged DF \n",
    "weekly_customer_fulldf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0116eb37",
   "metadata": {},
   "outputs": [],
   "source": [
    "## [OPTIONAL] pickle the DF:\n",
    "#weekly_customer_fulldf.to_pickle(\"./weekly_customer_fulldf.pkl\")\n",
    "#weekly_customer_fulldf = pd.read_pickle(\"./weekly_customer_fulldf.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f8b7b868",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b1d84e7a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5c1bd8c6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "253f6685",
   "metadata": {},
   "source": [
    "<a id='ide'></a>\n",
    "\n",
    "## INITIAL DATA EXPLORATION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8e4a3bfa",
   "metadata": {},
   "outputs": [],
   "source": [
    "weekly_customer_fulldf = pd.read_pickle(\"./weekly_customer_fulldf.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "3f070e11",
   "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>store</th>\n",
       "      <th>week_ending</th>\n",
       "      <th>sales</th>\n",
       "      <th>baskets</th>\n",
       "      <th>card_num</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>city</th>\n",
       "      <th>department</th>\n",
       "      <th>email_status</th>\n",
       "      <th>post_code</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.99</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>BREAD</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>5.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>CONFECTIONERY</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>DAIRY</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>15.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SEASONAL LINES</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SNACKFOODS</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  store   week_ending  sales  baskets        card_num    age  gender   city  \\\n",
       "0    MW  Apr 21, 2019   1.99     1.00  59626613925472 792.00  Female  Miami   \n",
       "1    MW  Apr 21, 2019   5.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "2    MW  Apr 21, 2019   1.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "3    MW  Apr 21, 2019  15.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "4    MW  Apr 21, 2019   3.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "\n",
       "       department  email_status  post_code  \n",
       "0           BREAD          True   4,220.00  \n",
       "1   CONFECTIONERY          True   4,220.00  \n",
       "2           DAIRY          True   4,220.00  \n",
       "3  SEASONAL LINES          True   4,220.00  \n",
       "4      SNACKFOODS          True   4,220.00  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#E.G. of data input for the visuals and tables\n",
    "weekly_customer_fulldf.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "14dcdce6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c3127c58",
   "metadata": {},
   "source": [
    "### BRIEF INTRODUCTION TO DEPARTMENT SALES DYNAMICS \n",
    "* Unsurprisingly, Tobacco has the highest average transaction value; although, this will likely not translate into significance in our classification model as tobacco does not count towards accumulating loyalty points. \n",
    "    * The same logic holds true for EPAY/PHONECARDS and LOTTO/SCRATCHIT\n",
    "    \n",
    "* Thus, the highest average value transaction for an actual department category is meat, followed by prized-poultry (cooked, ready-to-eat, BBQ meat out of the deli-case); these items generally have a fixed cost of 10 dollars so the average is also unsurprisingly around that amount. \n",
    "\n",
    "* We also observe fairly similar average spend amounts around other categories, like: fruit & veg, grocery, deli, dairy etc. We expect at least some of these items to be significant in the classification models "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "5b8847e2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(data=weekly_customer_fulldf, y='department', x='sales')\n",
    "plt.title('Department Average Transaction Value For All Customers Over Sample Period');\n",
    "# note: fuel_discount (will be dropped eventually) is negative because it is an actual discount on check-out value"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f72fa7b9",
   "metadata": {},
   "source": [
    "### DEPARTMENT TOTAL TRANSACTION SIZES\n",
    "* In terms of total number of transactions involving these departments, there seems to be a reasonable representation for all departments with the exception of: deli-retail and seasonal lines\n",
    "    * seasonal lines is somewhat expected given these items are only available at limited times of the year, however, they may play an important role in the cusotmer email-opt in preferences if people deem these limited-availability items of high demand, thus I will leave it in the model. \n",
    "    * deli-retail is an odd one, it actually represents a few specific deli lines, measured on how the items are sold (E.G. most are sold by weight, hence deli-items, but some are by fixed bundle sizes or units- these are deli-retail). More or less, the type of item is identical to what is found by deli-items, thus the two categories could be combined or deli-retail simply dropped due to it being almost a negligible quantity. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "dd3306fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(data=weekly_customer_fulldf, y='department', x='sales', estimator=len)\n",
    "plt.title('Department Total Transaction Counts For All Customers Over Sample Period');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1154d17c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "64c84b6e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "63ae0a43",
   "metadata": {},
   "source": [
    "\n",
    "## CUSTOMER AGE\n",
    "* This variable is partially \"corrupted\" - the values represent a scale; that is, smaller values = younger and larger values = older, thus it is not entirely useless. Although, I do not have access to the scale parameter to convert these values into real values that represent biological ages\n",
    "* At any rate, it can be observed that the bulk of the transactions (customers) are being made by people on the younger side of the scale\n",
    "    * This is potentially representative of the fact that younger people are more likely to adopt \"modern\" technology and be invested in customer loyalty / email marketing programs vs. their older counterparts etc.\n",
    "    \n",
    "    \n",
    "#### Important to note: age data is only available for 1/3 of the data set, meaning it cannot reliably make conclusions about the entire data set and we cannot apply CLT without first making an assumption about the distribution of the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "4b8bd01d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# definitely a skew towards the lower end of the age tier \n",
    "# this will be resolved if the variable is used in the final model by standardisation\n",
    "plt.hist(weekly_customer_fulldf.groupby(['card_num'])['age'].last())\n",
    "plt.title('Registered Age of Customer Associated with their Last Known Transaction');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d0ce6770",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c9cfb209",
   "metadata": {},
   "source": [
    "### BREAKDOWN OF AVERAGE SALES BY AGE SCALE"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71d9f8b3",
   "metadata": {},
   "source": [
    "* NOTE: Age is scaled and does not represent true values for privacy reasons when releasing the data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "fa5935e1",
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"5\" halign=\"left\">sales</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age</th>\n",
       "      <th>(-5.238, 5067.6]</th>\n",
       "      <th>(5067.6, 10115.2]</th>\n",
       "      <th>(10115.2, 15162.8]</th>\n",
       "      <th>(15162.8, 20210.4]</th>\n",
       "      <th>(20210.4, 25258.0]</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>email_status</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>7.22</td>\n",
       "      <td>11.23</td>\n",
       "      <td>7.20</td>\n",
       "      <td>6.79</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>7.85</td>\n",
       "      <td>11.08</td>\n",
       "      <td>17.97</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        sales                                       \\\n",
       "age          (-5.238, 5067.6] (5067.6, 10115.2] (10115.2, 15162.8]   \n",
       "email_status                                                         \n",
       "False                    7.22             11.23               7.20   \n",
       "True                     7.85             11.08              17.97   \n",
       "\n",
       "                                                    \n",
       "age          (15162.8, 20210.4] (20210.4, 25258.0]  \n",
       "email_status                                        \n",
       "False                      6.79                NaN  \n",
       "True                        NaN              11.19  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# show average sales by age bin (1/5) increments and by email_status\n",
    "pd.pivot_table(weekly_customer_fulldf, index='email_status', columns = pd.cut(weekly_customer_fulldf['age'], bins=5),\n",
    "              values=['sales'], aggfunc='mean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "c1839d2a",
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"5\" halign=\"left\">sales</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age</th>\n",
       "      <th>(-5.238, 5067.6]</th>\n",
       "      <th>(5067.6, 10115.2]</th>\n",
       "      <th>(10115.2, 15162.8]</th>\n",
       "      <th>(15162.8, 20210.4]</th>\n",
       "      <th>(20210.4, 25258.0]</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>email_status</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>2271</td>\n",
       "      <td>1352</td>\n",
       "      <td>309</td>\n",
       "      <td>248</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>5598</td>\n",
       "      <td>2795</td>\n",
       "      <td>1039</td>\n",
       "      <td>0</td>\n",
       "      <td>257</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        sales                                       \\\n",
       "age          (-5.238, 5067.6] (5067.6, 10115.2] (10115.2, 15162.8]   \n",
       "email_status                                                         \n",
       "False                    2271              1352                309   \n",
       "True                     5598              2795               1039   \n",
       "\n",
       "                                                    \n",
       "age          (15162.8, 20210.4] (20210.4, 25258.0]  \n",
       "email_status                                        \n",
       "False                       248                  0  \n",
       "True                          0                257  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# number of transactions in each age bin\n",
    "pd.pivot_table(weekly_customer_fulldf, index='email_status', columns = pd.cut(weekly_customer_fulldf['age'], bins=5),\n",
    "              values=['sales'], aggfunc='count')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b83ff70c",
   "metadata": {},
   "source": [
    "* Judging from the middle age cut (3/5) which by counts is actually refering to the roughly the oldest 10% of the customer base, the average spend is indeed significantly higher among email opt-ins vs. email opt-outs (17.97 Vs. 7.20) and across the last two bins (11.19 vs 6.79)\n",
    "    * This would suggest the older customer who chooses to opt into emails is also correlated with a higher spend; the direction of the causality (emails = more sales, or more sales already = opt into emails) cannot yet be infered."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0eec33ad",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "09b9c40c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c4c13490",
   "metadata": {},
   "source": [
    "## CUSTOMER GENDER\n",
    "* Gender is known for roughly 80% of the transactions in the data-set\n",
    "* Later, in feature engineering, a separate technique to handle the unknown genders will be introduced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "69dd4b01",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(weekly_customer_fulldf['gender'].value_counts().index,\n",
    "           weekly_customer_fulldf['gender'].value_counts()).set_title(\"Total Transactions by Gender\")\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5948e87b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ebe3633d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "a3e6c1f6",
   "metadata": {},
   "source": [
    "## CUSTOMER DYNAMICS BY EMAIL STATUS\n",
    "* The majority of transactions are made by people who have opted into emails; but there remains a significant portion of transactions by people with loyalty cards who have chosen not to. Thus, the interesting problem emerges of what underlying phenomena is there to explain this divide? "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "e8302713",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(weekly_customer_fulldf['email_status'].value_counts().index,\n",
    "           weekly_customer_fulldf['email_status'].value_counts()).set_title(\"Total Transactions by Email Status\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ac58be82",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "166d1f48",
   "metadata": {},
   "source": [
    "* The importance of converting people is demonstrated by the average sales value of customers who opt into email marketing being higher than those who don't; thus, it should be determined which factors can help convert these customers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "3d5173aa",
   "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>sales</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>email_status</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>8.70</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>10.04</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              sales\n",
       "email_status       \n",
       "False          8.70\n",
       "True          10.04"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.pivot_table(weekly_customer_fulldf, index='email_status',\n",
    "              values=['sales'], aggfunc='mean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "37ca45ae",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c4b44ef3",
   "metadata": {},
   "source": [
    "#### Further breaking down the average spend by email status with gender\n",
    "* We see a great deal of the gains in average spend stem from converting males to opting into email"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "7393e1d3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sales & Baskets Daily Transaction Counts By Gender & Email Status \n",
      " --------------------------------------------------\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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">sales</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gender</th>\n",
       "      <th>Female</th>\n",
       "      <th>Male</th>\n",
       "      <th>Unknown</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>email_status</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>9.60</td>\n",
       "      <td>6.62</td>\n",
       "      <td>9.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>9.53</td>\n",
       "      <td>10.42</td>\n",
       "      <td>10.37</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              sales              \n",
       "gender       Female  Male Unknown\n",
       "email_status                     \n",
       "False          9.60  6.62    9.10\n",
       "True           9.53 10.42   10.37"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Sales & Baskets Daily Transaction Counts By Gender & Email Status\", '\\n', \"--\"*25)\n",
    "\n",
    "pd.pivot_table(weekly_customer_fulldf, index='email_status', columns='gender',\n",
    "              values=['sales'], aggfunc='mean')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f3376ab9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a137de4c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "07238230",
   "metadata": {},
   "source": [
    "## Department Sales Dynamics By Email Status\n",
    "* The results are somewhat conflicting; the 3 main stay departments (tobacco, lotto, phonecards) have significantly higher average spends than their non-email counterparts; this might suggest that these items are staple/repeat purchases to customers who have already delegated the store as their chosen frequent store.\n",
    "* There are some modest gains in impulse categories (snackfoods, softdrinks) etc.\n",
    "    * although, these are generally balanced out by modest loses in other departments; resulting in a fairly neutral net outcome\n",
    "    \n",
    "    \n",
    "### NOTE: median was also performed and found virtually identical results; suggesting that the department sales distributions are not heavily skewed \n",
    "\n",
    "* Additionally, is it imperative to note that this data represents loyalty members (opt-in) vs loyalty members (email opt-out); a comparison between loyalty members in general vs. non-loyalty members average department spends would likely be more in line with expectations that loyalty members spend more than non-members. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "87d196a5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(rc={'figure.figsize':(14,8)})\n",
    "sns.barplot(x='department', y='sales', hue='email_status', data=weekly_customer_fulldf, estimator=np.mean)\n",
    "plt.xticks(rotation=70)\n",
    "plt.title('Department Sales Mean By Customer Email Status');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "21ffac4b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a13a7ddf",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "b068b028",
   "metadata": {},
   "source": [
    "## VISUALISE THE DISTRIBUTION OF INDIVIDUAL CUSTOMER'S (FIRST) TRANSACTION OVER TIME BY EMAIL STATUS\n",
    "\n",
    "* The distribution of customer first recorded loyalty transaction over the sample horizon, the data shows there are still a great deal of new customers entering the system, most of whom have opted into emails but some of whom still have chosen not to\n",
    "    * One interesting observation is the tendency for new transactions (akin to new sign-ups) clustering around December, particularly so in 2020 and 2021, meaning the holiday (summer) season is more attractive to signing up new customers\n",
    "    * Furthermore, quite a few customers have rather large first spend amounts, although the majority of customer's first transaction value is roughly 0-10 dollars; likely indicating they used it as a top-up or convenience shop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "f4f61273",
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert the DF time variable to datetime for the index of the graph\n",
    "weekly_customer_fulldf['week_ending'] = pd.to_datetime(weekly_customer_fulldf['week_ending'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "aaa0f04d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(rc={'figure.figsize':(14,8)})\n",
    "sns.scatterplot( data=weekly_customer_fulldf.groupby('card_num').first(),\n",
    "                y='sales', x='week_ending', hue='email_status')\n",
    "plt.title('Distribution of first transaction by customer & email status over sample horizon');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e9aed2c0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "8efb7ac0",
   "metadata": {},
   "source": [
    "### VISUALISE THE DISTRIBUTION OF INDIVIDUAL CUSTOMER'S (LAST) TRANSACTION OVER TIME BY EMAIL STATUS\n",
    "* Interesting, a lot of customers with email_status set to false made their last transactions around the same christmas time; this perhaps suggests they signed up for the holiday season and perhaps some factors contributed towards their decision not to return, or that their shopping needs/habits changed drastically after the holidays. \n",
    "\n",
    "* Interestingly, most customer's last transaction amount is less clustered around zero than their first transaction but there are far less extreme positive outliers here than in the first transaction group. (E.G. max transaction value here is ~ 65 dollars Vs. quite a few spends over 80 dollars for first transactions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "3e1aede3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#sns.set(rc={'figure.figsize':(14,8)})\n",
    "sns.scatterplot( data=weekly_customer_fulldf.groupby('card_num').last(),\n",
    "                y='sales', x='week_ending', hue='email_status')\n",
    "plt.title('Distribution of last transaction by customer & email status over sample horizon');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7495407d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c2ee9ad0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2014222a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "2832bb38",
   "metadata": {},
   "source": [
    "<a id='data_pre'></a>\n",
    "\n",
    "## DATA PREPROCESSING\n",
    "* Dummy Variable Creation For Departments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "d1259bd8",
   "metadata": {},
   "outputs": [],
   "source": [
    "## DF with department category aggregated (No Hot Encoding / Dummies)\n",
    "## OPTIONAL: load pickle from previous step\n",
    "weekly_customer_fulldf = pd.read_pickle(\"./weekly_customer_fulldf.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "425c823a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# create department-labeled dummy variables for each department\n",
    "dep_dummies = pd.get_dummies(weekly_customer_fulldf.department)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "7b960a55",
   "metadata": {},
   "outputs": [],
   "source": [
    "# there are a few specialised categories with their own labels \n",
    "# E.G. tobacco, health and beauty goods, lotto, magazines: simply re-name them for convenience\n",
    "dep_dummies.rename(columns={'CIGS & TOBACCO':'TOBACCO',\n",
    "                                         'HEALTH & BEAUTY':'H_AND_B',\n",
    "                                         'LOTTO/SCRATCHIT':'LOTTO',\n",
    "                           'NEWSPAPER/MAGS':'MAGAZINES'}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "7604ea77",
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge onto the main DF with the data\n",
    "weekly_customer_dep_dummies = pd.concat([weekly_customer_fulldf, \n",
    "                                        dep_dummies], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "bd425dd5",
   "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>store</th>\n",
       "      <th>week_ending</th>\n",
       "      <th>sales</th>\n",
       "      <th>baskets</th>\n",
       "      <th>card_num</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>city</th>\n",
       "      <th>department</th>\n",
       "      <th>email_status</th>\n",
       "      <th>post_code</th>\n",
       "      <th>BREAD</th>\n",
       "      <th>TOBACCO</th>\n",
       "      <th>CONFECTIONERY</th>\n",
       "      <th>DAIRY</th>\n",
       "      <th>DELI-ITEMS</th>\n",
       "      <th>DELI-RETAIL</th>\n",
       "      <th>EPAY/PHONECARD</th>\n",
       "      <th>FROZEN</th>\n",
       "      <th>FRUIT &amp; VEG</th>\n",
       "      <th>FUEL DISCOUNT</th>\n",
       "      <th>GENERAL MERCH</th>\n",
       "      <th>GROCERY</th>\n",
       "      <th>H_AND_B</th>\n",
       "      <th>INSTORE BAKERY</th>\n",
       "      <th>LAUNDRY NEEDS</th>\n",
       "      <th>LENARDS</th>\n",
       "      <th>LOTTO</th>\n",
       "      <th>MEAT</th>\n",
       "      <th>MILK - FRESH</th>\n",
       "      <th>MAGAZINES</th>\n",
       "      <th>PET FOOD &amp; ACC</th>\n",
       "      <th>PRIZE POULTRY</th>\n",
       "      <th>SEASONAL LINES</th>\n",
       "      <th>SNACKFOODS</th>\n",
       "      <th>SOFT DRINKS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.99</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>BREAD</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>5.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>CONFECTIONERY</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>DAIRY</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>15.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SEASONAL LINES</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>MW</td>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>Female</td>\n",
       "      <td>Miami</td>\n",
       "      <td>SNACKFOODS</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  store   week_ending  sales  baskets        card_num    age  gender   city  \\\n",
       "0    MW  Apr 21, 2019   1.99     1.00  59626613925472 792.00  Female  Miami   \n",
       "1    MW  Apr 21, 2019   5.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "2    MW  Apr 21, 2019   1.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "3    MW  Apr 21, 2019  15.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "4    MW  Apr 21, 2019   3.00     1.00  59626613925472 792.00  Female  Miami   \n",
       "\n",
       "       department  email_status  post_code  BREAD  TOBACCO  CONFECTIONERY  \\\n",
       "0           BREAD          True   4,220.00      1        0              0   \n",
       "1   CONFECTIONERY          True   4,220.00      0        0              1   \n",
       "2           DAIRY          True   4,220.00      0        0              0   \n",
       "3  SEASONAL LINES          True   4,220.00      0        0              0   \n",
       "4      SNACKFOODS          True   4,220.00      0        0              0   \n",
       "\n",
       "   DAIRY  DELI-ITEMS  DELI-RETAIL  EPAY/PHONECARD  FROZEN  FRUIT & VEG  \\\n",
       "0      0           0            0               0       0            0   \n",
       "1      0           0            0               0       0            0   \n",
       "2      1           0            0               0       0            0   \n",
       "3      0           0            0               0       0            0   \n",
       "4      0           0            0               0       0            0   \n",
       "\n",
       "   FUEL DISCOUNT  GENERAL MERCH  GROCERY  H_AND_B  INSTORE BAKERY  \\\n",
       "0              0              0        0        0               0   \n",
       "1              0              0        0        0               0   \n",
       "2              0              0        0        0               0   \n",
       "3              0              0        0        0               0   \n",
       "4              0              0        0        0               0   \n",
       "\n",
       "   LAUNDRY NEEDS  LENARDS  LOTTO  MEAT  MILK - FRESH  MAGAZINES  \\\n",
       "0              0        0      0     0             0          0   \n",
       "1              0        0      0     0             0          0   \n",
       "2              0        0      0     0             0          0   \n",
       "3              0        0      0     0             0          0   \n",
       "4              0        0      0     0             0          0   \n",
       "\n",
       "   PET FOOD & ACC  PRIZE POULTRY  SEASONAL LINES  SNACKFOODS  SOFT DRINKS  \n",
       "0               0              0               0           0            0  \n",
       "1               0              0               0           0            0  \n",
       "2               0              0               0           0            0  \n",
       "3               0              0               1           0            0  \n",
       "4               0              0               0           1            0  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# E.G. of the resulting DF \n",
    "weekly_customer_dep_dummies.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4cfac602",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "09cd57ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "# [OPTIONAL] pickle the DF:\n",
    "#weekly_customer_dep_dummies.to_pickle(\"./weekly_customer_dep_dummies.pkl\")\n",
    "weekly_customer_dep_dummies = pd.read_pickle(\"./weekly_customer_dep_dummies.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "25d44d7a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dbb122ba",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a1729526",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "ac2c1501",
   "metadata": {},
   "source": [
    "<a id='feature_eng'></a>\n",
    "\n",
    "## FINAL FEATURE ENGINEERING & DATA MANIPULATION / CLEANING \n",
    "* E.G. \n",
    "    * Create a customer_duration feature (how long have they been a loyalty member)\n",
    "    * Aggregate transaction counts by department and scale the result\n",
    "    * Create a dummy variable for each gender + 3rd category for missing gender data\n",
    "    * drop unneeded variables after some iterative testing \n",
    "    \n",
    "#### POTENTIAL EXPANSION AT A LATER DATE: \n",
    "* Use NLP to infer gender from names to fill the gab of missing gender data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "81b74b7e",
   "metadata": {},
   "outputs": [],
   "source": [
    "## read in pickle if resuming \n",
    "weekly_customer_dep_dummies = pd.read_pickle(\"./weekly_customer_dep_dummies.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "5327c598",
   "metadata": {},
   "outputs": [],
   "source": [
    "## NOT USED IN THE FINAL MODEL AFTER ITERATION\n",
    "\n",
    "\n",
    "# convert sales to log to assist with the highly skewed - un-even distribution \n",
    "weekly_customer_dep_dummies['log_sales'] = np.log(weekly_customer_dep_dummies.sales+1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e3791113",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c7cf27bd",
   "metadata": {},
   "source": [
    "### CREATE TOTAL SALES BY CUSTOMER FEATURE "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "1b77317b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# for testing purposes, make a copy of DF thus far \n",
    "weekly_customer_dep_dummies_cleaned = weekly_customer_dep_dummies.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "4e1cda07",
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop NaN from Log-Sales, which also dropping NaN from sales\n",
    "weekly_customer_dep_dummies_cleaned.dropna(subset=['log_sales'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c8828c51",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "cefc554d",
   "metadata": {},
   "source": [
    "### DROP MISSING AGE VALUES (EXPERIMENTAL)\n",
    "* Initially excluded from the first iteration, but re-added in subsequent tests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "164d7190",
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop missing age values (roughly 25% of customer data)\n",
    "weekly_customer_dep_dummies_cleaned.dropna(subset=['age'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "60848c64",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "04814cbb",
   "metadata": {},
   "source": [
    "### VARIABLES TO BE DROPPED AS THEY'RE OBSOLETE / INEFFECTIVE \n",
    "* City = replaced by post_code \n",
    "* Store = all customers in this data file pertain to the same store\n",
    "* Department = replaced by individual department dummy variables\n",
    "* Baskets \n",
    "    * No longer has any meaning in the context of aggregated department frequencies\n",
    "\n",
    "    \n",
    "### DROP NAN ROWS FROM SOME VARIABLES: \n",
    "* Sales / Post Code \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "c3202b9e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop those variables\n",
    "weekly_customer_dep_dummies_cleaned = weekly_customer_dep_dummies_cleaned.drop(['city','store', \n",
    "                                                                                'department', 'baskets'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "5541ff08",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pre (post_code) drop DF len: 13868\n",
      "post (post_code) drop DF len: 13741\n"
     ]
    }
   ],
   "source": [
    "# drop the subset of customers with missing postcodes: (not many)\n",
    "print(\"pre (post_code) drop DF len: {}\".format(len(weekly_customer_dep_dummies_cleaned)))\n",
    "weekly_customer_dep_dummies_cleaned.dropna(subset=['post_code'], inplace=True)\n",
    "print(\"post (post_code) drop DF len: {}\".format(len(weekly_customer_dep_dummies_cleaned)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6fed4c73",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "6cb984e4",
   "metadata": {},
   "source": [
    "### Encode a dummy variable for gender (M/F) & A separate variable to indicate missing gender values\n",
    "* A potentially more sophisticated method of guessing gender based of a NLP model could be implemented at a later date "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "9da65e30",
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to string \n",
    "weekly_customer_dep_dummies_cleaned['gender'] = weekly_customer_dep_dummies_cleaned.gender.astype(str)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "6e820ff6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# if gender is unknown - set dummy to (1) else (0)\n",
    "weekly_customer_dep_dummies_cleaned['gender_unknown'] = weekly_customer_dep_dummies_cleaned['gender'].apply(\n",
    "lambda x: 1 if x=='Unknown' else 0)\n",
    "\n",
    "# in the male column: set == 1 if male; else 0 \n",
    "weekly_customer_dep_dummies_cleaned['male'] = weekly_customer_dep_dummies_cleaned['gender'].apply(\n",
    "lambda x: 1 if x=='Male' else 0)\n",
    "\n",
    "# repeat for females \n",
    "weekly_customer_dep_dummies_cleaned['female'] = weekly_customer_dep_dummies_cleaned['gender'].apply(\n",
    "lambda x: 1 if x=='Female' else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "f1826278",
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop the original gender column as it is now not needed\n",
    "weekly_customer_dep_dummies_cleaned = weekly_customer_dep_dummies_cleaned.drop(['gender'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "64179bbb",
   "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>week_ending</th>\n",
       "      <th>sales</th>\n",
       "      <th>card_num</th>\n",
       "      <th>age</th>\n",
       "      <th>email_status</th>\n",
       "      <th>post_code</th>\n",
       "      <th>BREAD</th>\n",
       "      <th>TOBACCO</th>\n",
       "      <th>CONFECTIONERY</th>\n",
       "      <th>DAIRY</th>\n",
       "      <th>DELI-ITEMS</th>\n",
       "      <th>DELI-RETAIL</th>\n",
       "      <th>EPAY/PHONECARD</th>\n",
       "      <th>FROZEN</th>\n",
       "      <th>FRUIT &amp; VEG</th>\n",
       "      <th>FUEL DISCOUNT</th>\n",
       "      <th>GENERAL MERCH</th>\n",
       "      <th>GROCERY</th>\n",
       "      <th>H_AND_B</th>\n",
       "      <th>INSTORE BAKERY</th>\n",
       "      <th>LAUNDRY NEEDS</th>\n",
       "      <th>LENARDS</th>\n",
       "      <th>LOTTO</th>\n",
       "      <th>MEAT</th>\n",
       "      <th>MILK - FRESH</th>\n",
       "      <th>MAGAZINES</th>\n",
       "      <th>PET FOOD &amp; ACC</th>\n",
       "      <th>PRIZE POULTRY</th>\n",
       "      <th>SEASONAL LINES</th>\n",
       "      <th>SNACKFOODS</th>\n",
       "      <th>SOFT DRINKS</th>\n",
       "      <th>log_sales</th>\n",
       "      <th>gender_unknown</th>\n",
       "      <th>male</th>\n",
       "      <th>female</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.99</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.10</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>5.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.79</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>1.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.69</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>15.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2.77</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Apr 21, 2019</td>\n",
       "      <td>3.00</td>\n",
       "      <td>59626613925472</td>\n",
       "      <td>792.00</td>\n",
       "      <td>True</td>\n",
       "      <td>4,220.00</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>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>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.39</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    week_ending  sales        card_num    age  email_status  post_code  BREAD  \\\n",
       "0  Apr 21, 2019   1.99  59626613925472 792.00          True   4,220.00      1   \n",
       "1  Apr 21, 2019   5.00  59626613925472 792.00          True   4,220.00      0   \n",
       "2  Apr 21, 2019   1.00  59626613925472 792.00          True   4,220.00      0   \n",
       "3  Apr 21, 2019  15.00  59626613925472 792.00          True   4,220.00      0   \n",
       "4  Apr 21, 2019   3.00  59626613925472 792.00          True   4,220.00      0   \n",
       "\n",
       "   TOBACCO  CONFECTIONERY  DAIRY  DELI-ITEMS  DELI-RETAIL  EPAY/PHONECARD  \\\n",
       "0        0              0      0           0            0               0   \n",
       "1        0              1      0           0            0               0   \n",
       "2        0              0      1           0            0               0   \n",
       "3        0              0      0           0            0               0   \n",
       "4        0              0      0           0            0               0   \n",
       "\n",
       "   FROZEN  FRUIT & VEG  FUEL DISCOUNT  GENERAL MERCH  GROCERY  H_AND_B  \\\n",
       "0       0            0              0              0        0        0   \n",
       "1       0            0              0              0        0        0   \n",
       "2       0            0              0              0        0        0   \n",
       "3       0            0              0              0        0        0   \n",
       "4       0            0              0              0        0        0   \n",
       "\n",
       "   INSTORE BAKERY  LAUNDRY NEEDS  LENARDS  LOTTO  MEAT  MILK - FRESH  \\\n",
       "0               0              0        0      0     0             0   \n",
       "1               0              0        0      0     0             0   \n",
       "2               0              0        0      0     0             0   \n",
       "3               0              0        0      0     0             0   \n",
       "4               0              0        0      0     0             0   \n",
       "\n",
       "   MAGAZINES  PET FOOD & ACC  PRIZE POULTRY  SEASONAL LINES  SNACKFOODS  \\\n",
       "0          0               0              0               0           0   \n",
       "1          0               0              0               0           0   \n",
       "2          0               0              0               0           0   \n",
       "3          0               0              0               1           0   \n",
       "4          0               0              0               0           1   \n",
       "\n",
       "   SOFT DRINKS  log_sales  gender_unknown  male  female  \n",
       "0            0       1.10               0     0       1  \n",
       "1            0       1.79               0     0       1  \n",
       "2            0       0.69               0     0       1  \n",
       "3            0       2.77               0     0       1  \n",
       "4            0       1.39               0     0       1  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#E.G: gender_unknown = (0) as gender is known -- then, gender = (1) for female, (0) for male\n",
    "weekly_customer_dep_dummies_cleaned.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f0ee3acd",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97da4203",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "02edf401",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "93e10e77",
   "metadata": {},
   "source": [
    "## ADDITIONAL FEATURES CREATION \n",
    "* Average Spend by Customer Transactions\n",
    "* Customer duration (last-first transaction dates)\n",
    "\n",
    "\n",
    "#### HANDLING THE CLASS IMBALANCE: \n",
    "* This will be addressed slightly below by upscaling the minority class (resampling with replacement to match the majority class)\n",
    "    * It was possible in some models E.G. SVC to handle this directly, but other models did not support it so it was manually implemented beforehand for consistency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e76c21ae",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "cfb1a20f",
   "metadata": {},
   "source": [
    "### extract the email status and post-code of the last data entry by customer (presuming post-code doesn't change through-out the sample)\n",
    "* Essentially, extract variables that are not expected to change (excluding email status, which might change but will be handled later)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "c2cdaeb4",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_email_postcode_df = weekly_customer_dep_dummies_cleaned.groupby(['card_num'])[\n",
    "    'email_status','post_code','gender_unknown','male','female','age'].last()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "1bd64c24",
   "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>email_status</th>\n",
       "      <th>post_code</th>\n",
       "      <th>gender_unknown</th>\n",
       "      <th>male</th>\n",
       "      <th>female</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>card_num</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2710201787225</th>\n",
       "      <td>False</td>\n",
       "      <td>4,227.00</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>25.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202677549</th>\n",
       "      <td>False</td>\n",
       "      <td>4,067.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>29.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202683021</th>\n",
       "      <td>True</td>\n",
       "      <td>4,069.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710203126152</th>\n",
       "      <td>True</td>\n",
       "      <td>4,169.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710300512261</th>\n",
       "      <td>False</td>\n",
       "      <td>4,218.00</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               email_status  post_code  gender_unknown  male  female   age\n",
       "card_num                                                                  \n",
       "2710201787225         False   4,227.00               0     1       0 25.00\n",
       "2710202677549         False   4,067.00               0     0       1 29.00\n",
       "2710202683021          True   4,069.00               0     0       1 30.00\n",
       "2710203126152          True   4,169.00               0     0       1 40.00\n",
       "2710300512261         False   4,218.00               0     1       0 20.00"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_email_postcode_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "86491075",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "074a0f79",
   "metadata": {},
   "source": [
    "### CREATE SUMMARY OF DEPARTMENT FREQUENCIES BY CUSTOMER \n",
    "* By customer, sum each departments sales values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "b971df11",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "170\n"
     ]
    }
   ],
   "source": [
    "cust_dep_sums = weekly_customer_dep_dummies_cleaned.groupby(['card_num'])['BREAD', 'TOBACCO', 'CONFECTIONERY',\n",
    "       'DAIRY', 'DELI-ITEMS', 'DELI-RETAIL', 'EPAY/PHONECARD', 'FROZEN',\n",
    "       'FRUIT & VEG', 'FUEL DISCOUNT', 'GENERAL MERCH', 'GROCERY', 'H_AND_B',\n",
    "       'INSTORE BAKERY', 'LAUNDRY NEEDS', 'LENARDS', 'LOTTO', 'MEAT',\n",
    "       'MILK - FRESH', 'MAGAZINES', 'PET FOOD & ACC', 'PRIZE POULTRY',\n",
    "       'SEASONAL LINES', 'SNACKFOODS', 'SOFT DRINKS'].sum()\n",
    "\n",
    "# unique customer rows in the data \n",
    "print(len(cust_dep_sums))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5fcbd1da",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c5529fd2",
   "metadata": {},
   "source": [
    "## CREATE CUSTOMER SALES FEATURES\n",
    "* Include: Total Spend & Average Spend\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "5251e722",
   "metadata": {},
   "outputs": [],
   "source": [
    "cust_avgspend = weekly_customer_dep_dummies_cleaned.groupby([\n",
    "    'card_num'])['sales'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "60a65772",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "21d91a63",
   "metadata": {},
   "source": [
    "## Create time-spent as customer feature \n",
    "* measured by the time between their first and last shops "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "94c3f22a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# if not already done, convert the week_ending to datetime format\n",
    "weekly_customer_dep_dummies_cleaned['week_ending'] = pd.to_datetime(weekly_customer_dep_dummies_cleaned['week_ending'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "a0b78348",
   "metadata": {},
   "outputs": [],
   "source": [
    "# create the duration variable as last transction - first transaction date by customer\n",
    "cust_duration = weekly_customer_dep_dummies_cleaned.groupby(['card_num'])['week_ending'].max() - weekly_customer_dep_dummies_cleaned.groupby(['card_num'])['week_ending'].min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "62fb356d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to DF and transpose to match dimensions of existing data\n",
    "customer_durations_df = pd.DataFrame([cust_duration]).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6afbb403",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "255a9dfb",
   "metadata": {},
   "source": [
    "### Need to convert the data series to float (drop the .days part of the variable for the classification models)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "738ec0b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert the days timedelta variable into hours \n",
    "customer_durations_df['week_ending'] = customer_durations_df['week_ending'] / np.timedelta64(1, 'h')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "58816a41",
   "metadata": {},
   "outputs": [],
   "source": [
    "# divide the now floated (hours) variable by 24 to get back to days\n",
    "customer_durations_df['week_ending'] = customer_durations_df['week_ending'] / 24"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "2c4ef43e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# rename variable for clarity \n",
    "customer_durations_df.rename(columns={'week_ending':'cust_duration'}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "7dfc43d2",
   "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>cust_duration</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>card_num</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2710201787225</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202677549</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202683021</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710203126152</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710300512261</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>512246819655144</th>\n",
       "      <td>336.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>542060132417800</th>\n",
       "      <td>343.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>761594497527545</th>\n",
       "      <td>329.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>894399176741730</th>\n",
       "      <td>343.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>937764028975188</th>\n",
       "      <td>350.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>170 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 cust_duration\n",
       "card_num                      \n",
       "2710201787225             0.00\n",
       "2710202677549             0.00\n",
       "2710202683021             0.00\n",
       "2710203126152             0.00\n",
       "2710300512261             0.00\n",
       "...                        ...\n",
       "512246819655144         336.00\n",
       "542060132417800         343.00\n",
       "761594497527545         329.00\n",
       "894399176741730         343.00\n",
       "937764028975188         350.00\n",
       "\n",
       "[170 rows x 1 columns]"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# E.G. card_num and customer duration in (days)\n",
    "customer_durations_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "51da117e",
   "metadata": {},
   "source": [
    "* customer duration == 0 would represent customers who made a single purchase on the day they signed up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d4a6084f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ea66bb85",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "6adf7e69",
   "metadata": {},
   "source": [
    "## Aggregate (concatenate) each of the individual data feature DF(S) into one final DF for the classification models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "2c9e7139",
   "metadata": {},
   "outputs": [],
   "source": [
    "# concatenate each of the individual aggregated (card_num) data features DFs into one\n",
    "customer_data_aggregated_df = pd.concat([customer_email_postcode_df, customer_durations_df,\n",
    "                                        cust_avgspend, cust_dep_sums], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "065ea436",
   "metadata": {},
   "outputs": [],
   "source": [
    "## pickle the unscaled DF [NOT USED AFTER ITERATION]\n",
    "\n",
    "#customer_data_aggregated_df.to_pickle(\"./customer_data_aggregated_df.pkl\")\n",
    "#customer_data_aggregated_df = pd.read_pickle(\"./customer_data_aggregated_df.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d04901d1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "1ddfe843",
   "metadata": {},
   "source": [
    "## SCALED DATA DF VERSION\n",
    "* Same as the above DF, except the aggregated department transaction frequencies will be standardised with sklearn's standard scaler; \n",
    "    * The mean_false = will be passed to preserve the sparse matrix nature of the data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "8b5af753",
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a copy of the existing DF to apply the scaling factor to\n",
    "customer_data_aggregated_df_scaled = customer_data_aggregated_df.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "b366c243",
   "metadata": {},
   "outputs": [],
   "source": [
    "# combine the departments which all belong to the aggregate department category \"deli\"\n",
    "customer_data_aggregated_df_scaled['DELI'] = customer_data_aggregated_df_scaled['DELI-ITEMS'] + customer_data_aggregated_df_scaled[\n",
    "    'DELI-RETAIL'] + customer_data_aggregated_df_scaled['PRIZE POULTRY'] + customer_data_aggregated_df_scaled['LENARDS']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "e6ad1b0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop the original frames \n",
    "customer_data_aggregated_df_scaled = customer_data_aggregated_df_scaled.drop([\n",
    "    'DELI-ITEMS','DELI-RETAIL','PRIZE POULTRY','LENARDS'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca48e202",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "7018d68a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# pass in \"with_mean\"=false, to account for the sparse matrix \n",
    "scale = StandardScaler(with_mean=False)\n",
    "\n",
    "# scale transform the aggregated department frequencies\n",
    "customer_data_aggregated_df_scaled[['BREAD', 'TOBACCO', 'CONFECTIONERY',\n",
    "       'DAIRY', 'EPAY/PHONECARD', 'FROZEN',\n",
    "       'FRUIT & VEG', 'FUEL DISCOUNT', 'GENERAL MERCH', 'GROCERY', 'H_AND_B',\n",
    "       'INSTORE BAKERY', 'LAUNDRY NEEDS', 'LOTTO', 'MEAT',\n",
    "       'MILK - FRESH', 'MAGAZINES', 'PET FOOD & ACC',\n",
    "       'SEASONAL LINES', 'SNACKFOODS', 'SOFT DRINKS', 'DELI']] = scale.fit_transform(\n",
    "    \n",
    "customer_data_aggregated_df_scaled[['BREAD', 'TOBACCO', 'CONFECTIONERY',\n",
    "       'DAIRY', 'EPAY/PHONECARD', 'FROZEN',\n",
    "       'FRUIT & VEG', 'FUEL DISCOUNT', 'GENERAL MERCH', 'GROCERY', 'H_AND_B',\n",
    "       'INSTORE BAKERY', 'LAUNDRY NEEDS', 'LOTTO', 'MEAT',\n",
    "       'MILK - FRESH', 'MAGAZINES', 'PET FOOD & ACC',\n",
    "       'SEASONAL LINES', 'SNACKFOODS', 'SOFT DRINKS','DELI']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "7a5fcd2a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# scale the customer duration and sales (experimented with: mean=False and mean=True)\n",
    "scale = StandardScaler(with_mean=False)\n",
    "\n",
    "customer_data_aggregated_df_scaled[['cust_duration','sales','age']] = scale.fit_transform(\n",
    "customer_data_aggregated_df_scaled[['cust_duration','sales','age']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6e5caf79",
   "metadata": {},
   "outputs": [],
   "source": [
    "# could try standardising age with scaler & mean=True in subsequent iterations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "6a735a32",
   "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>email_status</th>\n",
       "      <th>post_code</th>\n",
       "      <th>gender_unknown</th>\n",
       "      <th>male</th>\n",
       "      <th>female</th>\n",
       "      <th>age</th>\n",
       "      <th>cust_duration</th>\n",
       "      <th>sales</th>\n",
       "      <th>BREAD</th>\n",
       "      <th>TOBACCO</th>\n",
       "      <th>CONFECTIONERY</th>\n",
       "      <th>DAIRY</th>\n",
       "      <th>EPAY/PHONECARD</th>\n",
       "      <th>FROZEN</th>\n",
       "      <th>FRUIT &amp; VEG</th>\n",
       "      <th>FUEL DISCOUNT</th>\n",
       "      <th>GENERAL MERCH</th>\n",
       "      <th>GROCERY</th>\n",
       "      <th>H_AND_B</th>\n",
       "      <th>INSTORE BAKERY</th>\n",
       "      <th>LAUNDRY NEEDS</th>\n",
       "      <th>LOTTO</th>\n",
       "      <th>MEAT</th>\n",
       "      <th>MILK - FRESH</th>\n",
       "      <th>MAGAZINES</th>\n",
       "      <th>PET FOOD &amp; ACC</th>\n",
       "      <th>SEASONAL LINES</th>\n",
       "      <th>SNACKFOODS</th>\n",
       "      <th>SOFT DRINKS</th>\n",
       "      <th>DELI</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>card_num</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2710201787225</th>\n",
       "      <td>False</td>\n",
       "      <td>4,227.00</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.63</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202677549</th>\n",
       "      <td>False</td>\n",
       "      <td>4,067.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.75</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202683021</th>\n",
       "      <td>True</td>\n",
       "      <td>4,069.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.84</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.11</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710203126152</th>\n",
       "      <td>True</td>\n",
       "      <td>4,169.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.43</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710300512261</th>\n",
       "      <td>False</td>\n",
       "      <td>4,218.00</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.85</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.17</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               email_status  post_code  gender_unknown  male  female  age  \\\n",
       "card_num                                                                    \n",
       "2710201787225         False   4,227.00               0     1       0 0.01   \n",
       "2710202677549         False   4,067.00               0     0       1 0.01   \n",
       "2710202683021          True   4,069.00               0     0       1 0.01   \n",
       "2710203126152          True   4,169.00               0     0       1 0.01   \n",
       "2710300512261         False   4,218.00               0     1       0 0.01   \n",
       "\n",
       "               cust_duration  sales  BREAD  TOBACCO  CONFECTIONERY  DAIRY  \\\n",
       "card_num                                                                    \n",
       "2710201787225           0.00   0.63   0.00     0.00           0.12   0.00   \n",
       "2710202677549           0.00   0.75   0.00     0.00           0.12   0.00   \n",
       "2710202683021           0.00   1.84   0.00     0.00           0.00   0.11   \n",
       "2710203126152           0.00   1.43   0.00     0.00           0.12   0.00   \n",
       "2710300512261           0.00   0.85   0.00     0.00           0.00   0.00   \n",
       "\n",
       "               EPAY/PHONECARD  FROZEN  FRUIT & VEG  FUEL DISCOUNT  \\\n",
       "card_num                                                            \n",
       "2710201787225            0.00    0.00         0.09           0.00   \n",
       "2710202677549            0.00    0.00         0.09           0.00   \n",
       "2710202683021            0.00    0.15         0.00           0.00   \n",
       "2710203126152            0.00    0.00         0.00           0.00   \n",
       "2710300512261            0.00    0.00         0.09           0.00   \n",
       "\n",
       "               GENERAL MERCH  GROCERY  H_AND_B  INSTORE BAKERY  LAUNDRY NEEDS  \\\n",
       "card_num                                                                        \n",
       "2710201787225           0.00     0.09     0.00            0.00           0.00   \n",
       "2710202677549           0.00     0.00     0.00            0.00           0.00   \n",
       "2710202683021           0.00     0.09     0.00            0.00           0.00   \n",
       "2710203126152           0.00     0.00     0.26            0.20           0.00   \n",
       "2710300512261           0.00     0.00     0.00            0.00           0.00   \n",
       "\n",
       "               LOTTO  MEAT  MILK - FRESH  MAGAZINES  PET FOOD & ACC  \\\n",
       "card_num                                                              \n",
       "2710201787225   0.00  0.00          0.00       0.00            0.00   \n",
       "2710202677549   0.00  0.00          0.00       0.00            0.00   \n",
       "2710202683021   0.00  0.00          0.00       0.00            0.00   \n",
       "2710203126152   0.00  0.00          0.00       0.00            0.00   \n",
       "2710300512261   0.00  0.00          0.00       0.00            0.00   \n",
       "\n",
       "               SEASONAL LINES  SNACKFOODS  SOFT DRINKS  DELI  \n",
       "card_num                                                      \n",
       "2710201787225            0.00        0.00         0.00  0.00  \n",
       "2710202677549            0.00        0.00         0.10  0.00  \n",
       "2710202683021            0.00        0.00         0.10  0.00  \n",
       "2710203126152            0.00        0.00         0.00  0.00  \n",
       "2710300512261            0.00        0.17         0.00  0.00  "
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#E.G. of the scaled department variables but with the sparse matrix format preserved by the transformation\n",
    "customer_data_aggregated_df_scaled.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9968d602",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f57ec43e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "d3a8ecbe",
   "metadata": {},
   "source": [
    "## HOT ENCODING / DUMMY VARIABLE FOR POST-CODE\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "9c2233a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# scale the customer duration and sales (experimented with: mean=False and mean=True)\n",
    "#scale = StandardScaler(with_mean=False)\n",
    "\n",
    "#customer_data_aggregated_df_scaled[['post_code']] = scale.fit_transform(\n",
    "#customer_data_aggregated_df_scaled[['post_code']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "53f71b6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# create department-labeled dummy variables for each department\n",
    "postcode_dummies = pd.get_dummies(customer_data_aggregated_df_scaled.post_code)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "125e60d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge onto the main DF with the data\n",
    "weekly_customer_dep_dummies = pd.concat([weekly_customer_fulldf, \n",
    "                                        dep_dummies], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "46eb384a",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_data_aggregated_df_scaled_2 = pd.concat([customer_data_aggregated_df_scaled, postcode_dummies], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "0bfcb8f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# change name back to keep consistency; could refactor variable names later \n",
    "customer_data_aggregated_df_scaled = customer_data_aggregated_df_scaled_2.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1e2fddf2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "912b70b3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "93ea985e",
   "metadata": {},
   "source": [
    "## CREATE THE (y) VARIABLE TO BE PREDICTED: (email_status) \n",
    "* Eventually, drop this variable from the main DF into its own DF so that the data can predict the outcome variable "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "fd314b9b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define the series from the main DF\n",
    "y_emailstatus = customer_data_aggregated_df_scaled['email_status']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f9ba956b",
   "metadata": {},
   "source": [
    "#### Convert the boolean (True/False) to (1, 0)\n",
    "* WHERE:\n",
    "    * 1 = True \n",
    "    * 0 = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "78d8605d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# this can be achieved by * (1) for each element; \n",
    "y_emailstatus = y_emailstatus*1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "780c0cab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "unique customers in sample: 170\n"
     ]
    }
   ],
   "source": [
    "print(\"unique customers in sample: {}\".format(len(y_emailstatus)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "8ce5d568",
   "metadata": {},
   "outputs": [],
   "source": [
    "# rename the sales variable to more accurately reflect what it is now\n",
    "customer_data_aggregated_df_scaled = customer_data_aggregated_df_scaled.rename(columns={'sales':'avg_sales'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "9e89a738",
   "metadata": {},
   "outputs": [],
   "source": [
    "#(AFTER ITERATION) dropped deli-retail & realised \"fuel_discount\" should have already been dropped (neg values; no intuition)\n",
    "customer_data_aggregated_df_scaled.drop(['FUEL DISCOUNT'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "2617b969",
   "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>email_status</th>\n",
       "      <th>post_code</th>\n",
       "      <th>gender_unknown</th>\n",
       "      <th>male</th>\n",
       "      <th>female</th>\n",
       "      <th>age</th>\n",
       "      <th>cust_duration</th>\n",
       "      <th>avg_sales</th>\n",
       "      <th>BREAD</th>\n",
       "      <th>TOBACCO</th>\n",
       "      <th>CONFECTIONERY</th>\n",
       "      <th>DAIRY</th>\n",
       "      <th>EPAY/PHONECARD</th>\n",
       "      <th>FROZEN</th>\n",
       "      <th>FRUIT &amp; VEG</th>\n",
       "      <th>GENERAL MERCH</th>\n",
       "      <th>GROCERY</th>\n",
       "      <th>H_AND_B</th>\n",
       "      <th>INSTORE BAKERY</th>\n",
       "      <th>LAUNDRY NEEDS</th>\n",
       "      <th>LOTTO</th>\n",
       "      <th>MEAT</th>\n",
       "      <th>MILK - FRESH</th>\n",
       "      <th>MAGAZINES</th>\n",
       "      <th>PET FOOD &amp; ACC</th>\n",
       "      <th>SEASONAL LINES</th>\n",
       "      <th>SNACKFOODS</th>\n",
       "      <th>SOFT DRINKS</th>\n",
       "      <th>DELI</th>\n",
       "      <th>2486.0</th>\n",
       "      <th>2489.0</th>\n",
       "      <th>4067.0</th>\n",
       "      <th>4069.0</th>\n",
       "      <th>4169.0</th>\n",
       "      <th>4210.0</th>\n",
       "      <th>4211.0</th>\n",
       "      <th>4213.0</th>\n",
       "      <th>4214.0</th>\n",
       "      <th>4215.0</th>\n",
       "      <th>4218.0</th>\n",
       "      <th>4220.0</th>\n",
       "      <th>4226.0</th>\n",
       "      <th>4227.0</th>\n",
       "      <th>4361.0</th>\n",
       "      <th>4370.0</th>\n",
       "      <th>4380.0</th>\n",
       "      <th>4849.0</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>card_num</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2710201787225</th>\n",
       "      <td>False</td>\n",
       "      <td>4,227.00</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.63</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</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>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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202677549</th>\n",
       "      <td>False</td>\n",
       "      <td>4,067.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.75</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.00</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>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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710202683021</th>\n",
       "      <td>True</td>\n",
       "      <td>4,069.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.84</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.11</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.00</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>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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710203126152</th>\n",
       "      <td>True</td>\n",
       "      <td>4,169.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.43</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</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>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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2710300512261</th>\n",
       "      <td>False</td>\n",
       "      <td>4,218.00</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.85</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.17</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</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>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",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               email_status  post_code  gender_unknown  male  female  age  \\\n",
       "card_num                                                                    \n",
       "2710201787225         False   4,227.00               0     1       0 0.01   \n",
       "2710202677549         False   4,067.00               0     0       1 0.01   \n",
       "2710202683021          True   4,069.00               0     0       1 0.01   \n",
       "2710203126152          True   4,169.00               0     0       1 0.01   \n",
       "2710300512261         False   4,218.00               0     1       0 0.01   \n",
       "\n",
       "               cust_duration  avg_sales  BREAD  TOBACCO  CONFECTIONERY  DAIRY  \\\n",
       "card_num                                                                        \n",
       "2710201787225           0.00       0.63   0.00     0.00           0.12   0.00   \n",
       "2710202677549           0.00       0.75   0.00     0.00           0.12   0.00   \n",
       "2710202683021           0.00       1.84   0.00     0.00           0.00   0.11   \n",
       "2710203126152           0.00       1.43   0.00     0.00           0.12   0.00   \n",
       "2710300512261           0.00       0.85   0.00     0.00           0.00   0.00   \n",
       "\n",
       "               EPAY/PHONECARD  FROZEN  FRUIT & VEG  GENERAL MERCH  GROCERY  \\\n",
       "card_num                                                                     \n",
       "2710201787225            0.00    0.00         0.09           0.00     0.09   \n",
       "2710202677549            0.00    0.00         0.09           0.00     0.00   \n",
       "2710202683021            0.00    0.15         0.00           0.00     0.09   \n",
       "2710203126152            0.00    0.00         0.00           0.00     0.00   \n",
       "2710300512261            0.00    0.00         0.09           0.00     0.00   \n",
       "\n",
       "               H_AND_B  INSTORE BAKERY  LAUNDRY NEEDS  LOTTO  MEAT  \\\n",
       "card_num                                                             \n",
       "2710201787225     0.00            0.00           0.00   0.00  0.00   \n",
       "2710202677549     0.00            0.00           0.00   0.00  0.00   \n",
       "2710202683021     0.00            0.00           0.00   0.00  0.00   \n",
       "2710203126152     0.26            0.20           0.00   0.00  0.00   \n",
       "2710300512261     0.00            0.00           0.00   0.00  0.00   \n",
       "\n",
       "               MILK - FRESH  MAGAZINES  PET FOOD & ACC  SEASONAL LINES  \\\n",
       "card_num                                                                 \n",
       "2710201787225          0.00       0.00            0.00            0.00   \n",
       "2710202677549          0.00       0.00            0.00            0.00   \n",
       "2710202683021          0.00       0.00            0.00            0.00   \n",
       "2710203126152          0.00       0.00            0.00            0.00   \n",
       "2710300512261          0.00       0.00            0.00            0.00   \n",
       "\n",
       "               SNACKFOODS  SOFT DRINKS  DELI  2486.0  2489.0  4067.0  4069.0  \\\n",
       "card_num                                                                       \n",
       "2710201787225        0.00         0.00  0.00       0       0       0       0   \n",
       "2710202677549        0.00         0.10  0.00       0       0       1       0   \n",
       "2710202683021        0.00         0.10  0.00       0       0       0       1   \n",
       "2710203126152        0.00         0.00  0.00       0       0       0       0   \n",
       "2710300512261        0.17         0.00  0.00       0       0       0       0   \n",
       "\n",
       "               4169.0  4210.0  4211.0  4213.0  4214.0  4215.0  4218.0  4220.0  \\\n",
       "card_num                                                                        \n",
       "2710201787225       0       0       0       0       0       0       0       0   \n",
       "2710202677549       0       0       0       0       0       0       0       0   \n",
       "2710202683021       0       0       0       0       0       0       0       0   \n",
       "2710203126152       1       0       0       0       0       0       0       0   \n",
       "2710300512261       0       0       0       0       0       0       1       0   \n",
       "\n",
       "               4226.0  4227.0  4361.0  4370.0  4380.0  4849.0  \n",
       "card_num                                                       \n",
       "2710201787225       0       1       0       0       0       0  \n",
       "2710202677549       0       0       0       0       0       0  \n",
       "2710202683021       0       0       0       0       0       0  \n",
       "2710203126152       0       0       0       0       0       0  \n",
       "2710300512261       0       0       0       0       0       0  "
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# E.G. final DF: \n",
    "customer_data_aggregated_df_scaled.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "091c4436",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ebb2d85f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "12189b30",
   "metadata": {},
   "source": [
    "### pickle the featured-created DF for the classification models:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "cc995cfc",
   "metadata": {},
   "outputs": [],
   "source": [
    "# PICKLE OBJECTS\n",
    "\n",
    "#customer_data_aggregated_df_scaled.to_pickle(\"./customer_data_aggregated_df_scaled.pkl\")\n",
    "customer_data_aggregated_df_scaled = pd.read_pickle(\"./customer_data_aggregated_df_scaled.pkl\") \n",
    "#y_emailstatus.to_pickle(\"./y_emailstatus.pkl\")\n",
    "y_emailstatus = pd.read_pickle(\"./y_emailstatus.pkl\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "181292f5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a4b00c62",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "3e3ea806",
   "metadata": {},
   "source": [
    "## HANDLING THE EXTREME CLASS IMBALANCE PROBLEM: UP-SAMPLE MINORITY CLASS\n",
    "* Tried both up-sampling minority and down-sampling majority class; found the up-sampling method to be superior"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "2aa2f95f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import sklearns utils function for resampling with replacement\n",
    "from sklearn.utils import resample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "86c33a9b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define the y_series from the column in the DF\n",
    "customer_data_aggregated_df_scaled['y_emailstatus'] = y_emailstatus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "d915faec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    120\n",
       "0     50\n",
       "Name: email_status, dtype: int64"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 120 to the majority class (email=True) Vs. 50 (email=False)\n",
    "y_emailstatus.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "2e88df0f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# split the data into two separate DF based off majority and minority classes\n",
    "customer_majority = customer_data_aggregated_df_scaled[\n",
    "    customer_data_aggregated_df_scaled.y_emailstatus==1]\n",
    "\n",
    "customer_minority = customer_data_aggregated_df_scaled[\n",
    "    customer_data_aggregated_df_scaled.y_emailstatus==0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "f107f488",
   "metadata": {},
   "outputs": [],
   "source": [
    "# upsample the minority class with replacement to match the length of the majority class (120)\n",
    "customer_data_minority_upsampled = resample(customer_minority,\n",
    "                                                         replace=True,\n",
    "                                                         n_samples=y_emailstatus.value_counts()[1],\n",
    "                                                         random_state=123)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "1502acc1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# re-join the now upsampled minority class with the original majority class \n",
    "# should now be balanced (120:120)\n",
    "customer_majority_upsampled_df = pd.concat([customer_data_minority_upsampled, customer_majority])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "566f619a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# pickle DF [optional]\n",
    "\n",
    "#customer_majority_upsampled_df.to_pickle(\"./customer_majority_upsampled_df.pkl\")\n",
    "customer_majority_upsampled_df = pd.read_pickle(\"./customer_majority_upsampled_df.pkl\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "50d33d32",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f15efe19",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "90ea7345",
   "metadata": {},
   "source": [
    "## POST PICKLE IMPORT DROPS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "57a23b1e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# check the two classes sizes again\n",
    "y_emailstatus_upsampled = customer_majority_upsampled_df['y_emailstatus']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "d035ea79",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    120\n",
       "1    120\n",
       "Name: y_emailstatus, dtype: int64"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# we see the two classes are now balanced for the classification models\n",
    "y_emailstatus_upsampled.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "3543bc27",
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop the now encoded variable: post_code & the y_variable (email_status) (as it will be the predicted series)\n",
    "customer_majority_upsampled_df = customer_majority_upsampled_df.drop(['email_status','y_emailstatus', 'post_code'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f872e8d7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "9e908884",
   "metadata": {},
   "source": [
    "## CREATE THE TRAIN-TEST SPLIT FOR CLASSIFICATION MODELS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "df9b6b45",
   "metadata": {},
   "outputs": [],
   "source": [
    "#test-train split (OG)\n",
    "X_train, X_test, y_train, y_test = train_test_split(customer_majority_upsampled_df, \n",
    "                                                   y_emailstatus_upsampled, test_size = 0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "0dd2b6c3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train len: 192\n",
      "X_test len: 48\n",
      "Y_train len: 192\n",
      "Y_test len: 48\n"
     ]
    }
   ],
   "source": [
    "# DF DETAILS \n",
    "print(\"X_train len: {}\".format(len(X_train)))\n",
    "print(\"X_test len: {}\".format(len(X_test)))\n",
    "print(\"Y_train len: {}\".format(len(y_train)))\n",
    "print(\"Y_test len: {}\".format(len(y_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "004bfb1a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(192,)"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# univaraite series to be \"classified\" / predicted \n",
    "y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "13c9dce3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(192, 45)"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# number of features in the data to create the predictions (45) and based of 192 rows of customer data \n",
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e5cf8582",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "15bd4a3a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "7feef191",
   "metadata": {},
   "source": [
    "<a id='models'></a>\n",
    "\n",
    "# CLASSIFICATION MODELS \n",
    "\n",
    "#### BROAD PROCEDURE\n",
    "* Import packages\n",
    "* Before initial un-tuned model testing & cross-validation to get an overview \n",
    "* Tune models (Ex. Naive Bayes) \n",
    "* Make predictions on test set to get accuracy score of tuned models \n",
    "* Results visualisation & summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "f10cc74b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import model packages \n",
    "from sklearn.ensemble import VotingClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import tree\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.svm import SVC # support vector classifier\n",
    "from sklearn.metrics import roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4c25b606",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "9742ce72",
   "metadata": {},
   "source": [
    "# CROSS VALIDATION\n",
    "* Get a broader-view of the performance/accuracy of the different models before actually extracting the predictions from tuned models for real use \n",
    "\n",
    "### UN-TUNED MODELS TO TEST INITIALLY: \n",
    "* Naive Bayes\n",
    "    * Baseline classification prediction\n",
    "* Logistic Regression\n",
    "* Decision Tree Classifier\n",
    "* Random Forest Classifier \n",
    "* K-Nearest Neighbors \n",
    "* Support Vector Classifier \n",
    "* XGBoost Classifier \n",
    "* (Soft) Voting Classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "ab4b5da2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initalise dictionary to store and later, quickly summarise the results of a few different classifiers \n",
    "cv_results_dict = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1aaf04a7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "d5f8fd88",
   "metadata": {},
   "source": [
    "## NAIVE BAYES -- BASELINE CLASSIFICATION PREDICTION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "29319c33",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.56 0.54 0.58 0.58 0.58] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.5679\n"
     ]
    }
   ],
   "source": [
    "# initialise model instance \n",
    "gnb = GaussianNB()\n",
    "\n",
    "# pass in 5 cross-validations (randomly samples from training data and predicts on that model there)\n",
    "cv = cross_val_score(gnb, \n",
    "                     X_train, y_train, cv=5)\n",
    "\n",
    "# print cross-validation iterations & mean\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['Naive_Bayes'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d4360771",
   "metadata": {},
   "source": [
    "* The baseline naive-bayes prediction performs ok, all things considered. Nevertheless, there is lots of room for improvement to provide significant gains in classification accuracy with the more complicated models.\n",
    "* Also, quite a bit of variation depending on the cross-validation sub-sample, which wouldn't be ideal in practice"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ab1cf931",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "392d70db",
   "metadata": {},
   "source": [
    "## LOGISTIC REGRESSION "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "512adf26",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.62 0.56 0.68 0.47 0.71] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.6096\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(max_iter = 2500)\n",
    "cv = cross_val_score(lr, X_train, y_train, cv=5)\n",
    "\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['Logistic_Regression'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7eb30626",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "76e23c72",
   "metadata": {},
   "source": [
    "## DECISION TREE CLASSIFIER "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "52f46eaf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.79 0.79 0.74 0.79 0.71] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.7653\n"
     ]
    }
   ],
   "source": [
    "dt = tree.DecisionTreeClassifier(random_state = 1)\n",
    "cv = cross_val_score(dt, \n",
    "                     X_train, y_train,cv=5)\n",
    "\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['Decision_Tree'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c939c6ef",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "967cbf3c",
   "metadata": {},
   "source": [
    "## RANDOM FOREST CLASSIFIER"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "c0ab0b76",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.79 0.85 0.79 0.92 0.82] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.8335\n"
     ]
    }
   ],
   "source": [
    "rf = RandomForestClassifier(random_state = 1)\n",
    "cv = cross_val_score(rf,\n",
    "                     X_train, y_train,cv=5)\n",
    "\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['Random_Forest'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d0b2c38c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c698c12e",
   "metadata": {},
   "source": [
    "## K-NEAREST NEIGHBORS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "df7076fa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.62 0.54 0.58 0.55 0.61] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.5781\n"
     ]
    }
   ],
   "source": [
    "knn = KNeighborsClassifier()\n",
    "cv = cross_val_score(knn, \n",
    "                     X_train, y_train,cv=5)\n",
    "\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['K_nearest_n'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cb083f3f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "fe7568fa",
   "metadata": {},
   "source": [
    "## SUPPORT VECTOR CLASSIFIER "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "b23f40e5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.54 0.54 0.74 0.55 0.74] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.6206\n"
     ]
    }
   ],
   "source": [
    "svc = SVC(probability = True, class_weight=None)\n",
    "cv = cross_val_score(svc, \n",
    "                     X_train, y_train, cv=5)\n",
    "\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['Support_Vector_Classifier'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0513057d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "d7e7e785",
   "metadata": {},
   "source": [
    "## XGBOOST CLASSIFIER (UN-TUNED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "9e5298fe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[12:35:30] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:30] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:30] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:30] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:30] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[0.77 0.82 0.71 0.84 0.74] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.7758\n"
     ]
    }
   ],
   "source": [
    "xgb = XGBClassifier(random_state =1)\n",
    "cv = cross_val_score(xgb,\n",
    "                     X_train,y_train, cv=5)\n",
    "\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['XGBoost_Classifier'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58212306",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce229ba4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "89b368e8",
   "metadata": {},
   "source": [
    "## VOTING CLASSIFIER \n",
    "* Because naive-bayes won't be included, there are an even number of voting candidates (hence, we won't use \"hard\" voting, which gives each classifier a (yes/no); from which, the traditional approach of the popular vote wins is the underlying process\n",
    "* A soft classifier averages the confidence of each model\n",
    "    * Accordingly, if the average of the models is > 50%, this will be assigned a (1), else assigned a (0) \n",
    "    * Whereby, the average might be model (1) = 95% and model (2) = 30%, thus average = (95 + 30) / 2  - which is > 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "cb3bf2f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define the voting classifier (execluding naive bayes)\n",
    "voting_clf = VotingClassifier(estimators = [\n",
    "    ('lr',lr),('knn',knn),('rf',rf),('gnb',gnb),('svc',svc),('xgb',xgb)], \n",
    "                              voting ='soft')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "id": "9a60cb2b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[12:35:36] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:37] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:37] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:37] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[12:35:37] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[0.77 0.59 0.63 0.74 0.71] \n",
      " ---------------------------------------- \n",
      " CV MEAN:  0.6876\n"
     ]
    }
   ],
   "source": [
    "cv = cross_val_score(voting_clf,\n",
    "                     X_train, y_train,cv=5)\n",
    "print(cv, '\\n', '--'*20, '\\n', \"CV MEAN: \", np.round(cv.mean(),4))\n",
    "cv_results_dict['Voting_Classifier'] = np.round(cv.mean(),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2f4d7937",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2e603cbe",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "8cb0e794",
   "metadata": {},
   "source": [
    "## CROSS-VALIDATION (PRE-TUNED) CLASSIFICATION MODELS COMPARISON"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "042d82ba",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Naive_Bayes': 0.5679,\n",
       " 'Logistic_Regression': 0.6096,\n",
       " 'Decision_Tree': 0.7653,\n",
       " 'Random_Forest': 0.8335,\n",
       " 'K_nearest_n': 0.5781,\n",
       " 'Support_Vector_Classifier': 0.6206,\n",
       " 'XGBoost_Classifier': 0.7758,\n",
       " 'Voting_Classifier': 0.6876}"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_results_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "87fc8e27",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show values in bar graph \n",
    "plt.figure(figsize=(10, 7))\n",
    "keys = cv_results_dict.keys()\n",
    "values = cv_results_dict.values()\n",
    "plt.bar(keys, values)\n",
    "plt.xticks(rotation='45')\n",
    "plt.title(\"Pre-Tuned Classificaiton CV Model Performances (Training Data)\");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3735693d",
   "metadata": {},
   "source": [
    "## BRIEF RESULTS SUMMARY\n",
    "\n",
    "* Definitively shows that the more complex models return significant accuracy improvements over the baseline naive bayes.\n",
    "* The RF seems to perform quite well without additional tuning; the XGBoost underperforms but we can expect fairly significant accuracy gains for XGBoost and the SVC in tuning. \n",
    "* The simplier logistic_regression model doesn't outperform the baseline naive bayes by much, suggesting that investing in more sophisticated models will be worthwhile\n",
    "    \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "13dea4be",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93ce1964",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "88d1b62b",
   "metadata": {},
   "source": [
    "<a id='tuned_models'></a>\n",
    "\n",
    "# TUNED CLASSIFICATION MODEL PREDICTIONS \n",
    "* Tune the above models (Excluding Naive Bayes) and re-test the performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "5ce080e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import additional packages for grid-searching to optimise some models \n",
    "from sklearn.model_selection import GridSearchCV, RandomizedSearchCV\n",
    "\n",
    "# pickle models to avoid having to re-run lengthy optimisation processes [optional]\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "924f7aa4",
   "metadata": {},
   "outputs": [],
   "source": [
    "## quick and easy print function for each model \n",
    "def model_performance_report(classifier, model_name):\n",
    "    print('\\n', \"MODEL: \", model_name, '\\n', '--'*15, '\\n', )\n",
    "    print('Best Score: ' + str(classifier.best_score_))\n",
    "    print('Best Parameters: ' + str(classifier.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "2014083b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialise model_tuned performance dictionary\n",
    "cv_tuned_results_dict = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a00216c5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7cb8532d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "3f1f4857",
   "metadata": {},
   "source": [
    "# TUNED LOGISTIC REGRESSION "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "cd1cd59b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "\n",
      " MODEL:  Logistic Regression \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.6354925775978407\n",
      "Best Parameters: {'C': 4.281332398719396, 'max_iter': 2500, 'penalty': 'l2', 'solver': 'liblinear'}\n"
     ]
    }
   ],
   "source": [
    "# initialise logistic_regression instance\n",
    "lr = LogisticRegression()\n",
    "\n",
    "# initialise param_grid of parameters to be optimised in the log-regression\n",
    "## note: logspace = numbers spaced evenly on a log scale\n",
    "param_grid = {'max_iter' : [2500],\n",
    "              'penalty' : ['l1', 'l2'],\n",
    "              'C' : np.logspace(-4, 4, 20),\n",
    "              'solver' : ['liblinear']}\n",
    "\n",
    "\n",
    "# create the grid-search object CV, passing in the parameters to be optimsied from the param_grid\n",
    "# note: n_jobs = -1 (will use all processors of your computer)\n",
    "clf_lr = GridSearchCV(lr, param_grid=param_grid, \n",
    "                      cv = 5, \n",
    "                      verbose =True, \n",
    "                      n_jobs = -1)\n",
    "\n",
    "# fit model passing in the grid-search object & our data: X_train and y_train\n",
    "best_clf_lr = clf_lr.fit(X_train, y_train)\n",
    "\n",
    "# pass in the auxiliary performance reporting function for the \"best\" / optimised parameter model chosen by grid-search\n",
    "model_performance_report(best_clf_lr,'Logistic Regression')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "c304f0f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "## append results to dictionary \n",
    "cv_tuned_results_dict['tuned_logreg'] = np.round(best_clf_lr.best_score_,4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "31957016",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "da487406",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c4e66840",
   "metadata": {},
   "source": [
    "## TUNED K-NEAREST NEIGHBORS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "1f05765f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 60 candidates, totalling 300 fits\n",
      "\n",
      " MODEL:  KNN \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.7342780026990554\n",
      "Best Parameters: {'algorithm': 'auto', 'n_neighbors': 3, 'p': 2, 'weights': 'distance'}\n"
     ]
    }
   ],
   "source": [
    "knn = KNeighborsClassifier()\n",
    "\n",
    "# KNN PARAM GRID\n",
    "param_grid = {'n_neighbors' : [3,5,7,9,11],\n",
    "              'weights' : ['uniform', 'distance'],\n",
    "              'algorithm' : ['auto', 'ball_tree','kd_tree'],\n",
    "              'p' : [1,2]}\n",
    "\n",
    "# run the KNN model with the grid_search passed in\n",
    "clf_knn = GridSearchCV(knn, param_grid=param_grid, cv=5, \n",
    "                       verbose=True, n_jobs = -1)\n",
    "\n",
    "# fit grid_searched model \n",
    "best_clf_knn = clf_knn.fit(X_train, y_train)\n",
    "\n",
    "# print brief report\n",
    "model_performance_report(best_clf_knn,'KNN')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "c624a137",
   "metadata": {},
   "outputs": [],
   "source": [
    "## append results to dictionary \n",
    "cv_tuned_results_dict['tuned_knn'] = np.round(best_clf_knn.best_score_,4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f5da2b7a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d6987b9f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "f42cb403",
   "metadata": {},
   "source": [
    "## TUNED SUPPORT VECTOR CLASSIFIER\n",
    "* PARAM: C = default (1.0); strength of the regularisation parameter "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dc0f03f3",
   "metadata": {},
   "source": [
    "#### NARROW DOWN FEATURE SPACE WITH RANDOMIZED SEARCH \n",
    "* SVC can have massive storage requirements, which increase greatly with the number of features; even sparse matrix will utilise lots of memory and the solver used in the algorithm doesn't scale efficiently with large amounts of sample\n",
    "* There are generally a few options if you are unable to perform the optimisation:\n",
    "    * Potentially try using linearSVC, which is vastly more efficient in larger sample sizes\n",
    "    * If data is sparse, replace (n) features with the average number of non-zero features in the sample vector\n",
    "    * down-scale your original majority class to match the minority class (reducing the size of the data)\n",
    "    * Tune the optimisation parameters (reduce max_iter) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "id": "0b6f7b78",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n",
      "\n",
      " MODEL:  SVC_rndCV \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.9008097165991904\n",
      "Best Parameters: {'kernel': 'rbf', 'gamma': 10, 'C': 1}\n"
     ]
    }
   ],
   "source": [
    "# fit the SVC with probability = True (internally uses 5-fold cross-validation)\n",
    "svc = SVC(probability = True, cache_size=2000, max_iter=500)\n",
    "# this max_iter is a hard_limit on iterations within solver\n",
    "# default cache=200mb, which is too small to test many different iterations of models\n",
    "\n",
    "# initialise param grid for randomized SVC params\n",
    "param_grid = tuned_parameters = [{'kernel': ['rbf'], 'gamma': [.1,.5,1,2,5,10],\n",
    "                                  'C': [.1, 1, 10, 100]},\n",
    "                                 {'kernel': ['linear'], 'C': [.1, 1, 10, 100]},\n",
    "                                 {'kernel': ['poly'], 'degree' : [2,3,4,5], 'C': [.1, 1, 10]}]\n",
    "\n",
    "\n",
    "# run CVC model with 20 parameter settings sampled\n",
    "clf_svc_rnd = RandomizedSearchCV(svc, param_distributions=param_grid, \n",
    "                       cv= 5, verbose= True, n_jobs= -1, n_iter=40)\n",
    "\n",
    "# fit the best model \n",
    "best_clf_svc_rnd = clf_svc_rnd.fit(X_train, y_train)\n",
    "\n",
    "# print brief report\n",
    "model_performance_report(best_clf_svc_rnd,'SVC_rndCV')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f9c8d8fc",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "089da79d",
   "metadata": {},
   "source": [
    "#### OPTIMIZE AROUND THE BEST PARAMETER FROM THE N_ITER(40) / CV=5,  RANDOMIZED SEARCH\n",
    "* Although, it seems like the model isn't improved by further refinement anyhow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "d6b2398e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 24 candidates, totalling 120 fits\n",
      "\n",
      " MODEL:  SVC \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.9008097165991904\n",
      "Best Parameters: {'C': 0.6, 'gamma': 10, 'kernel': 'rbf'}\n"
     ]
    }
   ],
   "source": [
    "# fit the SVC with probability = True (internally uses 5-fold cross-validation)\n",
    "svc = SVC(probability = True, cache_size=2000, max_iter=500)\n",
    "\n",
    "# initialise param grid for SVC params\n",
    "param_grid = tuned_parameters = [{'kernel': ['rbf'], 'gamma': [10, 15, 20],\n",
    "                                  'C': [0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.25, 2]}]\n",
    "\n",
    "# run CVC model\n",
    "clf_svc = GridSearchCV(svc, param_grid=param_grid, \n",
    "                       cv= 5, verbose= True, n_jobs= -1)\n",
    "\n",
    "# fit the best model \n",
    "best_clf_svc = clf_svc.fit(X_train, y_train)\n",
    "\n",
    "# print brief report\n",
    "model_performance_report(best_clf_svc,'SVC')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "c26c9126",
   "metadata": {},
   "outputs": [],
   "source": [
    "## append results to dictionary \n",
    "cv_tuned_results_dict['tuned_svc'] = np.round(best_clf_svc.best_score_,4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "b27179f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model [optional]\n",
    "pickle.dump(best_clf_svc, open(\"svc_best.pickle.dat\", \"wb\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8c0eac2e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b2e4de0d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "78394987",
   "metadata": {},
   "source": [
    "## TUNED RANDOM FOREST CLASSIFIER\n",
    "   \n",
    "#### Because this data set has a very large total feature space which can be very computationally intensive \n",
    "* first, perform a randomized search of the params, then do a more refined search from those params"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "a6b3bff1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 100 candidates, totalling 500 fits\n",
      "\n",
      " MODEL:  Random Forest \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.8439946018893387\n",
      "Best Parameters: {'n_estimators': 1000, 'min_samples_split': 2, 'min_samples_leaf': 1, 'max_features': 'log2', 'max_depth': None, 'bootstrap': False}\n"
     ]
    }
   ],
   "source": [
    "rf = RandomForestClassifier(random_state = 1)\n",
    "\n",
    "# initialise param grid for random_forest\n",
    "# note: default criterion is \"gini\" \n",
    "param_grid =  {'n_estimators': [100,500,1000], \n",
    "                                  'bootstrap': [True,False],\n",
    "                                  'max_depth': [3,5,10,20,50,75,100,None],\n",
    "                                  'max_features': ['auto','sqrt','log2'],\n",
    "                                  'min_samples_leaf': [1,2,4,10],\n",
    "                                  'min_samples_split': [2,5,10]}\n",
    "    \n",
    "## perform a randomized search of the params, then do a more refined search from those params\n",
    "clf_rf_rnd = RandomizedSearchCV(rf, param_distributions = param_grid, \n",
    "                                n_iter = 100, cv = 5, verbose = True, n_jobs = -1)\n",
    "\n",
    "# fit and print brief summary\n",
    "best_clf_rf_rnd = clf_rf_rnd.fit(X_train, y_train)\n",
    "model_performance_report(best_clf_rf_rnd,'Random Forest')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "883ae786",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "291fe33a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "995690dc",
   "metadata": {},
   "source": [
    "## REFINED GRID-SEARCH FOR TUNED RANDOM FOREST CLASSIFIER "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "5e3d0cc3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 36 candidates, totalling 180 fits\n",
      "\n",
      " MODEL:  Random Forest \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.8492577597840756\n",
      "Best Parameters: {'bootstrap': False, 'criterion': 'gini', 'max_depth': None, 'max_features': 'log2', 'min_samples_leaf': 1, 'min_samples_split': 3, 'n_estimators': 1000}\n"
     ]
    }
   ],
   "source": [
    "rf = RandomForestClassifier(random_state = 1)\n",
    "\n",
    "# pass in a range of values roughly around the Randomized Approximation values from above\n",
    "param_grid =  {'n_estimators': [800,1000,1200],\n",
    "               'criterion':['gini','entropy'],\n",
    "                                  'bootstrap': [False],\n",
    "                                  'max_depth': [None],\n",
    "                                  'max_features': ['log2'],\n",
    "                                  'min_samples_leaf': [1,2],\n",
    "                                  'min_samples_split': [1,2,3]}\n",
    " \n",
    "# fit the grid-search object\n",
    "clf_rf = GridSearchCV(rf, param_grid = param_grid, \n",
    "                      cv = 5, verbose = True, n_jobs = -1)\n",
    "\n",
    "best_clf_rf = clf_rf.fit(X_train, y_train)\n",
    "model_performance_report(clf_rf,'Random Forest')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "c872f150",
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_tuned_results_dict['tuned_rfc'] = np.round(best_clf_rf.best_score_,4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "c30472ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model \n",
    "pickle.dump(best_clf_rf, open(\"rf_best.pickle.dat\", \"wb\"))\n",
    "\n",
    "# load model \n",
    "#best_clf_rf = pickle.load(open(\"rf_best.pickle.dat\", \"rb\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4140a789",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f95e257c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "efe90808",
   "metadata": {},
   "source": [
    "### SHOW FEATURE IMPORTANCES IN RF\n",
    "* visualise the contribution of the 30 largest features in the RF classification algoirithm."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "7b85ade7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA/UAAAHRCAYAAAArGZnwAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAB74UlEQVR4nOzdeXiNd/7/8efJThJLCKGNCmopo2hLh0HF0tJaGnuI2qooFWtSaxAkttRSsVRoQyyxjKVpp1WdVilt+bZarSr9NZOIJYglQdbz+8M409PYE45zez2ua67LuZfP/f7ceX/nO69zf845JrPZbEZERERERERE7I6DrQsQERERERERkXujUC8iIiIiIiJipxTqRUREREREROyUQr2IiIiIiIiInVKoFxEREREREbFTCvUiIiIiIiIidsrJ1gXI7eXk5JKWdtnWZYgUWMmSRdXLYvfUx2IE6mMxAvWxGMH1Pvb29rznMfSk3g44OTnaugSRQqFeFiNQH4sRqI/FCNTHYgSF0ccK9SIiIiIiIiJ2SqFeRERERERExE7pM/U3cfbsWQICAoiJiSErK4upU6fi6OiIi4sLkZGRlC5dGoBz587RrVs3tm3bhqurK5cvX2bkyJFcuHCBIkWKMGvWLLy8vKzGXr9+PWvXrsXJyYlBgwbRrFmzW9bSduSW+zZPW4sJ9bd1CSIiIiIiInZLT+pvIDs7m4kTJ+Lm5gbAtGnTmDBhArGxsbRs2ZJly5YBsGvXLvr27cuZM2cs565fv56aNWsSFxfHyy+/zKJFi6zGTk1NJTY2lrVr17J8+XLmzp1LVlbWg5uciIiIiIiIGIZC/Q1ERkbSrVs3ypQpA8DcuXOpUaMGALm5ubi6ugLg4ODAihUrKFGihOXc3r17M2jQIABSUlIsT/SvO3jwIHXr1sXFxQVPT08qVKjA4cOHH8CsRERERERExGi0/P4vNm3ahJeXF40bN2bp0qUAlnB/4MABVq1axerVqwFo1KjRDcdwdHSkV69eHDlyhBUrVljtS09Px9Pzfz9X4O7uTnp6+v2Yil0oyE83iH3S31yMQH0sRqA+FiNQH4sRFLSPFer/YuPGjZhMJr7++mt++eUXQkJCiI6O5ttvvyU6OpqlS5fm+4z8jXzwwQccO3aMN954gx07dli2e3h4kJGRYXmdkZFhFfIfNampl2xdgjxA3t6e+puL3VMfixGoj8UI1MdiBNf7uCDBXqH+L64/hQcICgoiLCyMPXv2sG7dOmJjY62W2t/IkiVLKFu2LB06dKBo0aI4Olr/7mDt2rV55513yMzMJCsri2PHjlG1atX7MRURERERERExOIX628jLy2PatGmUK1eOoUOHAvDcc8/x1ltv3fD4jh07EhISwsaNG8nNzWX69OkArFixggoVKtC8eXOCgoIIDAzEbDYzfPhwy2f0RURERERERO6GyWw2m21dhNyelhaJEWiZnBiB+liMQH0sRqA+FiMojOX3+vZ7ERERERERETulUC8iIiIiIiJipxTqRUREREREROyUQr2IiIiIiIiInVKoFxEREREREbFTCvUiIiIiIiIidkq/U38TZ8+eJSAggJiYGLKyspg6dSqOjo64uLgQGRlJ6dKlWblyJR9++CEATZs2ZciQIeTm5jJjxgx++uknsrKyGDp0KM2aNbMae/369axduxYnJycGDRqUb/9ftR255b7N82EWE+pv6xJEREREREQeagr1N5Cdnc3EiRNxc3MDYNq0aUyYMIEaNWqwdu1ali1bRs+ePdm6dSvx8fGYTCYCAwNp0aIFP//8Mzk5Oaxdu5ZTp07x0UcfWY2dmppKbGwsGzduJDMzk8DAQBo1aoSLi4stpioiIiIiIiJ2TMvvbyAyMpJu3bpRpkwZAObOnUuNGjUAyM3NxdXVFR8fH9577z0cHR1xcHAgJycHV1dXvvrqK3x8fBgwYADjx4/H39/6afPBgwepW7cuLi4ueHp6UqFCBQ4fPvzA5ygiIiIiIiL2T6H+LzZt2oSXlxeNGze2bLse7g8cOMCqVavo3bs3zs7OeHl5YTabiYyM5KmnnsLPz4+0tDQSExNZsmQJr7/+Om+//bbV+Onp6Xh6elpeu7u7k56e/mAmJyIiIiIiIoai5fd/sXHjRkwmE19//TW//PILISEhREdH8+233xIdHc3SpUvx8vICIDMzk7Fjx+Lu7s6kSZMAKFGiBC+88AImk4n69evzxx9/WI3v4eFBRkaG5XVGRoZVyJf/8fbWfTEi/V3FCNTHYgTqYzEC9bEYQUH7WKH+L1avXm35d1BQEGFhYezZs4d169YRGxtLiRIlADCbzQwePJgGDRowYMAAyznPPPMMX3zxBS+++CKHDx+mXLlyVuPXrl2bd955h8zMTLKysjh27BhVq1Z9IHOzN6mpl2xdghQyb29P/V3F7qmPxQjUx2IE6mMxgut9XJBgr1B/G3l5eUybNo1y5coxdOhQAJ577jlq1KjBN998Q1ZWFrt27QJgxIgRdOnShUmTJtGlSxfMZjOTJ08GYMWKFVSoUIHmzZsTFBREYGAgZrOZ4cOH4+rqarP5iYiIiIiIiP0ymc1ms62LkNvTu5BiBHpHXYxAfSxGoD4WI1AfixEUxpN6fVGeiIiIiIiIiJ1SqBcRERERERGxUwr1IiIiIiIiInZKoV5ERERERETETinUi4iIiIiIiNgphXoRERERERERO6VQLyIiIiIiImKnnG53wL59+wgODqZKlSqWbSVLlqRo0aIcOnSIEiVKAJCbm8vkyZN58sknyczMxN/fnz59+tC/f3927txJVFQUGzduxMXFBYAZM2bg7OzMqFGjOHfuHLNmzaJ8+fJs376dMmXKAJCdnc3w4cNp0KAB/v7+fPTRR7i6ugJw7NgxwsLCiI2NxWw2ExcXx/bt23Fyujal/v3707RpUwCqVavGu+++S4sWLQD48ssvSUhIICIigqCgIK5cuUKRIkUs8+vXrx8vvPACJ06cICIignPnznH16lVq1qzJ2LFjLXP44Ycf6NGjB3FxcdSuXRuATZs2MX/+fHx9fQG4ePEi9erVY9KkSVb30mw2k5OTQ69evWjTps0t/wZtR2653Z/pkRUT6m/rEkRERERERGzmtqEe4PnnnycqKspqW2hoKKNHj6ZJkyYAfPHFF8ybN4+FCxfyr3/9izZt2rB582b69u2Lv78/O3bsYNGiRQQHB3PgwAH279/PmjVrLOc2adKEo0eP0rt3b7p37w5cC+6jRo1i8+bNt6xv3bp1HDhwgJUrV+Lq6kpaWhoDBgygePHi1KlThyJFihAREUG9evXw8vLKd35kZCSVK1e22pabm8vgwYMJCwvj6aefBiA8PJz58+czatQoAOLj4+nTp49VqAd45ZVXLMfk5eURGBjIjz/+mO9eZmRkEBQUhJ+fHzVq1LiDv4SIiIiIiIjI/xTa8vsLFy5QtGhR4FrY7dixI9WrV+eLL74AYOzYsSQkJHDo0CHCw8OJjIzE2dkZgK+++orGjRvnG/P8+fOWMW9l1apVjBs3zvIUv2TJkgwZMsTypoG7uzt9+vQhLCzsjuezf/9+fHx8LIEeYPTo0bz55pvAtUC+d+9ehgwZwoEDBzh37twNx8nIyODSpUt4enrm2+fu7k7Xrl35+OOP77guERERERERkevu6En93r17CQoKsry+vqx91qxZLFu2DAcHB8qUKcPo0aP5448/uHLlCtWrV6djx47ExMTQrFkzPDw8CA8P57XXXmPkyJGWJ+PZ2dlkZmbi4eEBwMqVK0lISMDBwYFixYoxdepUy3X79u2Lg8O19yH+vGQ+LS0t3xN4X19fUlJSLK8DAwP57LPP2LZtG8WLF7c6NiQkxGr5/bx58zh9+rRlCf111980AEhISKBly5a4urrSunVrNmzYwIABAwDYvn0733//Pampqbi7uzNw4EAqVqzIqVOn8t3bUqVKcejQoVvef7k5b+/8b5bIw01/MzEC9bEYgfpYjEB9LEZQ0D4utOX3182aNYsrV67Qr18/AA4cOEBiYiJPPPEE9evXp1ixYgQEBFiO/+6773j22Wctr/+8/P6vYmJi8n2mHsDDw4Pz589bPt8PkJiYSLly5SyvTSYT06dPp0ePHgwaNMhq3Bstvy9fvjyffPKJ1ba0tDS+//57mjVrRnx8PI6OjvTr14+rV69y8uRJ+vfvD/xv+X1SUhL9+/enYsWKN5wPQEpKCj4+PjfdL7eWmnrJ1iXIXfD29tTfTOye+liMQH0sRqA+FiO43scFCfaF+u33OTk5JCQksHr1apYvX87y5csZMGAAcXFxNz3n888/54UXXijQdXv27El4eDhZWVkAnD17loULF9KtWzer43x8fBg6dChz5sy57Zh16tQhOTmZgwcPAmA2m1m4cCHffvstv/76K7m5uaxZs4bly5ezevVqKlSowOeff241hq+vL5MmTWLYsGFcuXIl3zXS09OJj4/npZdeutepi4iIiIiIyCPsnpbfw7Vl43+1c+dOatasafXEPCAggPbt2xMcHGy1xP26xMTEWz7JvhNBQUHk5ubSo0cPnJycMJlMDB48mHr16uU7tkOHDnz66adW2/66/L5169YEBgYyb948pkyZwpUrV7h8+TJ16tQhODiYmTNn0r59e6sxOnfuzOrVq3nllVestjds2JCGDRsyf/58XnjhBcu9dHBwIDc3l6FDh1KpUqUCzV9EREREREQeTSaz2Wy2dRFye1paJEagZXJiBOpjMQL1sRiB+liM4KFbfi8iIiIiIiIiD45CvYiIiIiIiIidUqgXERERERERsVMK9SIiIiIiIiJ2SqFeRERERERExE4p1IuIiIiIiIjYqTv6nfpHSW5uLgMGDODy5cssXryY4sWLF8q4jRo1Yvfu3fd0btuRWwqlhkdFTKi/rUsQERERERF5IBTq/yI1NZW0tDQ2bdpk61JEREREREREbkmh/i8mTJjAH3/8wdtvv01GRgZpaWkAjB8/nmrVqtGyZUvq1q1LYmIizz//PJcuXeLgwYP4+fkxa9Ysjhw5QkREBHl5eVy8eJHx48dTr149y/i//vor4eHhAJQoUYLp06fj6elpk7mKiIiIiIiIfdNn6v9i0qRJVKlSBS8vL55//nliY2OZOnUqYWFhABw/fpzg4GBWrVrFBx98QGBgIPHx8ezfv5+LFy9y9OhRQkJCWLlyJX369Mn3xH/ChAlMmjSJ2NhYmjRpwnvvvWeDWYqIiIiIiIgR6En9TRw5coS9e/fy0UcfAXDx4kXg2tP18uXLA1C0aFGqVKkCgKenJ5mZmZQpU4ZFixbh5uZGRkYGHh4eVuMeO3aMyZMnA5CdnY2fn9+DmtIjw9tbKx8eZvr7iBGoj8UI1MdiBOpjMYKC9rFC/U1UqlSJdu3a0bZtW86ePUt8fDwAJpPpludNmzaN2bNnU7lyZebPn8/x48et9vv5+REZGUn58uXZv38/qamp920Oj6rU1Eu2LkFuwtvbU38fsXvqYzEC9bEYgfpYjOB6Hxck2CvU38TAgQMZN24c69evJz09nSFDhtzRee3atWPw4MGUKlUKHx8fy2fyrwsLCyMkJITc3Fzg2psAIiIiIiIiIvfCZDabzbYuQm5P70KKEegddTEC9bEYgfpYjEB9LEZQGE/q9UV5IiIiIiIiInZKoV5ERERERETETinUi4iIiIiIiNgphXoRERERERERO6VQLyIiIiIiImKnFOpFRERERERE7JRCvYiIiIiIiIidcrJ1AY+KTZs28fvvvzNq1Ki7PrftyC33oaJHS0yov61LEBERERERKXR6Ui8iIiIiIiJip/Sk/h5s2rSJzz//nKtXr5KamkqvXr347LPP+O233xgzZgwnT57kk08+IScnB09PTxYsWGB1fmxsLNu3b8dkMtGmTRt69eplo5mIiIiIiIiIPVOov0cZGRnExMTw4YcfsnLlStavX8++fftYuXIltWrVYuXKlTg4ONCvXz9+/PFHy3lHjx4lISGBuLg4TCYTvXv35h//+AeVKlWy4WyMz9vb09YlyH/pbyFGoD4WI1AfixGoj8UICtrHCvX3qEaNGgB4enpSuXJlTCYTxYsXJzs7G2dnZ0aMGEHRokU5efIkOTk5lvOOHDlCSkoKvXv3BuDChQv85z//Uai/z1JTL9m6BOHaf2HpbyH2Tn0sRqA+FiNQH4sRXO/jggR7hfp7ZDKZbrg9OzubHTt2EB8fz5UrVwgICMBsNlv2V6pUiSpVqvDee+9hMplYuXIlVatWfVBli4iIiIiIiIEo1BcyJycnihQpQkBAAC4uLnh7e3P69GnL/urVq/P3v/+d7t27k5WVRe3atSlbtqwNKxYRERERERF7ZTL/+TGyPLS0tEiMQMvkxAjUx2IE6mMxAvWxGEFhLL/XT9qJiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETslEK9iIiIiIiIiJ1SqBcRERERERGxU/qd+ps4e/YsAQEBxMTEkJWVxdSpU3F0dMTFxYXIyEhKly4NwLlz5+jWrRvbtm3D1dWVS5cuMXz4cK5cuYKzszOzZs3C29vbauz169ezdu1anJycGDRoEM2aNbtlLW1Hbrlv85Q7FxPqb+sSRERERERErOhJ/Q1kZ2czceJE3NzcAJg2bRoTJkwgNjaWli1bsmzZMgB27dpF3759OXPmjOXcTZs2UbVqVVavXk2bNm1Yvny51dipqanExsaydu1ali9fzty5c8nKynpwkxMRERERERHDUKi/gcjISLp160aZMmUAmDt3LjVq1AAgNzcXV1dXABwcHFixYgUlSpSwnFu1alUyMjIASE9Px8nJejHEwYMHqVu3Li4uLnh6elKhQgUOHz78AGYlIiIiIiIiRqNQ/xebNm3Cy8uLxo0bW7ZdD/cHDhxg1apV9O7dG4BGjRpRsmRJq/NLlizJ7t27LU/pO3XqZLU/PT0dT09Py2t3d3fS09Pv02xERERERETEyPSZ+r/YuHEjJpOJr7/+ml9++YWQkBCio6P59ttviY6OZunSpXh5ed30/IULF9K/f3+6devG4cOHGTp0KNu2bbPs9/DwsDzJB8jIyLAK+fLw8vbW36kw6D6KEaiPxQjUx2IE6mMxgoL2sUL9X6xevdry76CgIMLCwtizZw/r1q0jNjbWaqn9jRQrVswS0kuVKmUV4AFq167NO++8Q2ZmJllZWRw7doyqVasW+jyk8KWmXrJ1CXbP29tT91HsnvpYjEB9LEagPhYjuN7HBQn2CvW3kZeXx7Rp0yhXrhxDhw4F4LnnnuOtt9664fHDhg1j/PjxxMXFkZOTw9SpUwFYsWIFFSpUoHnz5gQFBREYGIjZbGb48OGWz+iLiIiIiIiI3A2T2Ww227oIuT29CylGoHfUxQjUx2IE6mMxAvWxGEFhPKnXF+WJiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETslEK9iIiIiIiIiJ1SqBcRERERERGxUwr1IiIiIiIiInbK6V5P3LdvH2vXriUqKirfvszMTPz9/enTpw/9+/cHIDk5mREjRrB+/XrLcWvWrOHMmTMMHTqUatWq8e6779KiRQsAvvzySxISEoiIiCAoKIgrV65QpEgRsrOzefzxxxk3bhwlS5YkNDSUQ4cOUaJECcxmM+fPn6dPnz40atSIzp07s3r1aipUqADAzp07Wbp0KatXr8bR0RGATZs2sXDhQrZu3YqHhwcAw4cPp1u3bjz22GO0a9eOmjVrWs1v5cqVLFq0iO3bt1OmTBnL9oYNGzJo0CD8/f0pV64cDg4OZGZmUrNmTUJDQ3F1dSUxMZFp06aRm5tLTk4OtWrVYuTIkTg43Pz9lbYjt9zLn0jsTEyov61LEBERERERO3PPof5W/vWvf9GmTRs2b95M3759bxlYrytSpAgRERHUq1cPLy+vfPsjIyOpXLkyAFu3bmXixIksWLAAgNGjR9OkSRMAzp8/zyuvvEJAQAAjR45k7NixxMbGcvHiRWbOnMmyZcssgf66K1euMH36dKZPn57vulWqVCE2NvaGNffu3Zvu3bvfcF9MTAyurq4AREdHExUVRWhoKHPnzqVnz540adIEs9nMkCFD+Oyzz2jZsuVt75GIiIiIiIjIn92X5ffx8fF07NiR6tWr88UXX9zROe7u7vTp04ewsLDbHtuuXTsOHTpEZmZmvn1nzpzBxcUFk8lEhw4dKFmyJOvWrSMyMpKBAwfi6+ub75wOHTrw+++/8/nnn99RrXerT58+fPLJJwCUL1+ezZs3s3//fnJycnjnnXcsqxNERERERERE7kahP6n/448/uHLlCtWrV6djx47ExMTQrFmzmx5vMpks/w4MDOSzzz5j27ZtFC9e/JbXKVasGBcvXgRg1qxZLF68mJSUFCpXrsy8efMsx02ePJmuXbvyt7/9jQ4dOtxwLEdHRyIiInj99depU6eO1b6jR48SFBRkeX19KT1cW4afkJBg2Tdw4EAaNWqUb3w3NzfLGxDDhw8nLi6OuXPncuTIEZo2bcrEiRMpVqzYLecrxuft7WnrEh6IR2WeYmzqYzEC9bEYgfpYjKCgfVzooT4+Pp4rV67Qr18/AA4cOEBiYiLu7u5kZWVZHXv58mXLEnW4FvCnT59Ojx49GDRo0E2vYTabOXPmDKVKlQL+t/z+iy++YPbs2ZbP0AN4eXnxzDPP0KZNm1vWXbFiRXr16sXkyZOt3mi41+X3f5aeno67uzsAe/fupXfv3vTu3ZuMjAwiIyNZtGiR5Y0CeXSlpl6ydQn3nbe35yMxTzE29bEYgfpYjEB9LEZwvY8LEuwLdfl9Tk4OCQkJrF69muXLl7N8+XIGDBhAXFwcpUqVIiMjg6NHjwKQm5vLnj17+Nvf/mY1ho+PD0OHDmXOnDk3vc6GDRt4/vnn831Wv2nTpjRv3pwJEybcU/09e/bk/Pnz7N27957Ov5lly5bRunVr4Nqqgt27dwPXPnLg5+eHi4tLoV5PREREREREHg0FelK/e/duAgICLK9bt25NzZo1KVGihGVbQEAA7du3Jzg4mBkzZjB27FgcHBzIzs6mefPmPP/88/nG7dChA59++qnVtpCQEIoUKQJA2bJlmTRp0g1rGjx4MAEBAfz73//mhRdeuKv5XF8p0LZtW8u2vy6/ByxfqPfX5fd+fn5MmTIFwPIFgXl5edSoUYMxY8YA8M477xAeHs6cOXNwcXHh8ccfv6PvERARERERERH5K5PZbDbbugi5PS0tEiPQMjkxAvWxGIH6WIxAfSxG8NAtvxcRERERERGRB0ehXkRERERERMROKdSLiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETsVIF+p37p0qXs2bMHBwcHTCYTw4cPp1atWoSGhnLo0CGr36tv164dnTt3BuCHH36gR48exMXFUbt2bQDy8vKIjIzkyJEjODg44OzszLhx4/D19cVsNhMXF8f27dtxcrpWcv/+/WnatCkA1apV491336VFixYAfPnllyQkJBAREQFAZmYm/v7+9OnTh/79+wOQnJzMiBEjWL9+/Q3ntm/fPtauXUtUVJTV9tDQUNq0aUOlSpV48cUXWbduHbVq1QJgzZo1nDlzhqFDh+Lv70+5cuVwcPjf+yYhISHUqlXrpvftZtqO3HJnfxB55MWE+tu6BBEREREReYDuOdQfPXqUnTt3smbNGkwmE7/88gshISFs3boVgNGjR9OkSZMbnhsfH0+fPn2sQv2uXbs4ffo0K1asAGDHjh1Mnz6d6Oho1q1bx4EDB1i5ciWurq6kpaUxYMAAihcvTp06dShSpAgRERHUq1cPLy+vfNf717/+RZs2bdi8eTN9+/a1CtoF4eHhwdtvv83GjRtxcXHJtz8mJgZXV1erbbe7byIiIiIiIiJ36p7TrZeXFykpKWzYsIFTp05Ro0YNNmzYcNvzMjIy2Lt3L0OGDOHAgQOcO3cOAB8fH3766ScSEhI4d+4czZs3Z968eQCsWrWKcePGWQJyyZIlGTJkCGvWrAHA3d2dPn36EBYWdsNrxsfH07FjR6pXr84XX3xxr1PO54knnqBx48b5nubfyr3eNxEREREREZG/KlCoj46O5sCBA3Tt2pWXXnqJzz//3LJ/1qxZBAUFWf7z66+/ApCQkEDLli1xdXWldevWlkBbrVo1pk6dyo4dO3jllVfo2LEj33//PQBpaWn5nsD7+vqSkpJieR0YGEh6ejrbtm2zOu6PP/7gypUrVK9enY4dO7J69ep7nfINBQcHs3v3br777rt8+/r27WuZ/2uvvQbc/r6JiIiIiIiI3Kl7Xn6fmJiIh4cHM2bMAODHH39kwIABNGjQALj58vv4+HgcHR3p168fV69e5eTJk/Tv358jR47g5+fH3LlzMZvN7N692xKYPTw8OH/+vNVn9BMTEylXrpzltclkYvr06fTo0YNBgwZZXe/KlSv069cPgAMHDpCYmIijo+O9Tt2Ki4sLM2bMYOTIkXTp0sVq342W39/qvv15fiL3wtvb09Yl3JY91ChyO+pjMQL1sRiB+liMoKB9fM+h/tdff2XNmjUsXrwYV1dX/Pz88PT0vGVY/vXXX8nNzbX6cro+ffrw+eef85///IfDhw8zffp0HB0defLJJylSpAgmk4mePXsSHh7O9OnTcXFx4ezZsyxcuJBx48ZZje/j48PQoUOJjIykadOm5OTkkJCQwObNmy2BOTo6mri4OIKCgu516vnUrFmTV155hWXLlhEYGHjLY+/lvoncqdTUS7Yu4Za8vT0f+hpFbkd9LEagPhYjUB+LEVzv44IE+3sO9a1ateLYsWN07tyZokWLYjabGTNmDJ6e14qZNWsWy5Ytsxz/3HPPcfHiRdq3b281TufOnVm9ejVLly4lMjKSDh064OHhgYODAzNnzgQgKCiI3NxcevTogZOTEyaTicGDB1OvXr18dXXo0IFPP/0UgJ07d1KzZk2rJ+ABAQG0b9+ezp0789tvvxEQEGDZFxoaSv369S2vd+/ebbV/zpw5N70fAwcOzLeM/q9fyterV6/b3jcRERERERGRO2Uym81mWxcht6d3IcUI9I66GIH6WIxAfSxGoD4WIyiMJ/WF89tuIiIiIiIiIvLAKdSLiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETslEK9iIiIiIiIiJ1SqBcRERERERGxU062LqCwREREcOjQIVJTU7l69Sq+vr6ULFmS0aNHM3PmTM6fP092djbVq1dn1KhReHh4sGDBArZv306ZMmXIycmhVKlSRERE4OHhAcCpU6do1aoVERERtG7d2nKt3377jVmzZnHlyhUuX75M06ZNGTp0KCaTiRMnThAREcG5c+e4evUqNWvWZOzYsbi4uJCRkUFUVBS//PILDg4OuLu7ExISgp+f3y3n1nbklvt678T4YkL9bV2CiIiIiIjcByaz2Wy2dRGFadOmTfz++++MGjWKq1ev0rlzZ8LDw3n66acB2Lx5Mx9//DFLlixhwYIFlC5dmu7duwMwd+5cihcvTr9+/QCIjo7m8uXLfP/998TGxgJw8eJFevTowYIFC6hYsSK5ubkMGzaMRo0a0aVLFzp16kRYWJjleuHh4bi5uTFq1ChGjBhB3bp1CQoKAuDw4cOMGDGCdevW4enpedM5KdRLQT0sod7b25PU1Eu2LkOkQNTHYgTqYzEC9bEYwfU+9va+eR68HUMvv//3v//Nc889ZwnYAK+++ippaWkkJSXlO/7ChQs89thjAJjNZrZs2UKfPn3Izs7myJEjAHz22Wc0aNCAihUrAuDo6EhkZCQdO3Zk//79+Pj4WF1v9OjRvPnmm5w7d44jR45YAj1A9erVadasGZ988sn9mL6IiIiIiIgYnGGW399IUlISFSpUyLf98ccfJyUlBYCVK1eSkJDA+fPnuXz5MoMHDwbg66+/pmrVqnh5edGxY0dWr17N5MmTOX36NL6+vlbjubu7A9xwn6urK3Btyf5f9wH4+vpaahG5Xwryzl9he5hqEblX6mMxAvWxGIH6WIygoH1s6FBftmxZDh48mG/7H3/8Qfny5QHo3bu3Zfl9fHw8ISEhrFy5kvXr15OcnEy/fv3Izs7m8OHDjBo1ivLly/Pzzz9bjZeUlMTJkycpX758vqfuaWlpfP/999SoUeOG4T0xMZHKlSsX1pRFbuhhWZqmZXJiBOpjMQL1sRiB+liMQMvvb6N58+bs2bPHKtjHx8fj5eV1w6fm5cuXJzs7m3PnzvHDDz8QHx/P8uXL+eCDD2jVqhWbN2+mWbNm7Nq1i//85z8AZGdnExERwZEjR6hTpw7JycmW65nNZhYuXMi3336Lj48PFSpUYPXq1ZbrHTp0iJ07d9KqVav7fCdERERERETEiAz9pN7d3Z3Fixczffp0zp8/T25uLtWqVWPu3LmWY64vv3d0dOTq1auMHTuWLVu20KpVKxwdHS3HdenShTFjxhAUFERERATjx4/HbDaTkZFBs2bNCAwMxGQyMW/ePKZMmWL5Zvw6deoQHBwMQGRkJDNnzqRz5844OjpSrFgxFi1aRLFixR70rREREREREREDMNy33xuVlhaJEWiZnBiB+liMQH0sRqA+FiPQ8nsRERERERGRR5hCvYiIiIiIiIidUqgXERERERERsVMK9SIiIiIiIiJ2SqFeRERERERExE4p1IuIiIiIiIjYKUP/Tn1BnD17loCAAGJiYsjKymLq1Kk4Ojri4uJCZGQkpUuXZuXKlXz44YcANG3alCFDhnD16lVGjx7N2bNncXd3JzIyEi8vL6ux169fz9q1a3FycmLQoEE0a9bslrW0Hbnlvs1THh0xof62LkFERERERAqZntTfQHZ2NhMnTsTNzQ2AadOmMWHCBGJjY2nZsiXLli0jKSmJrVu3snbtWtatW8dXX33F4cOHWbNmDVWrViUuLo4OHTqwaNEiq7FTU1OJjY1l7dq1LF++nLlz55KVlWWLaYqIiIiIiIidU6i/gcjISLp160aZMmUAmDt3LjVq1AAgNzcXV1dXfHx8eO+993B0dMTBwYGcnBxcXV3Zv38/jRs3BqBJkyZ8/fXXVmMfPHiQunXr4uLigqenJxUqVODw4cMPdoIiIiIiIiJiCAr1f7Fp0ya8vLwswRywhPsDBw6watUqevfujbOzM15eXpjNZiIjI3nqqafw8/MjPT0dT09PANzd3bl06ZLV+H/ef/2Y9PT0BzAzERERERERMRp9pv4vNm7ciMlk4uuvv+aXX34hJCSE6Ohovv32W6Kjo1m6dKnlM/KZmZmMHTsWd3d3Jk2aBICHhwcZGRkAZGRkUKxYMavx/7z/+jF/Dvki94u398PRZw9LHSIFoT4WI1AfixGoj8UICtrHCvV/sXr1asu/g4KCCAsLY8+ePaxbt47Y2FhKlCgBgNlsZvDgwTRo0IABAwZYzqlXrx5ffPEFtWvX5ssvv+SZZ56xGr927dq88847ZGZmkpWVxbFjx6hateoDmZs82lJTL93+oPvM29vzoahDpCDUx2IE6mMxAvWxGMH1Pi5IsFeov428vDymTZtGuXLlGDp0KADPPfccNWrU4JtvviErK4tdu3YBMGLECLp3705ISAjdu3fH2dmZOXPmALBixQoqVKhA8+bNCQoKIjAwELPZzPDhw3F1dbXZ/ERERERERMR+mcxms9nWRcjt6V1IMQK9oy5GoD4WI1AfixGoj8UICuNJvb4oT0RERERERMROKdSLiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETslEK9iIiIiIiIiJ1SqBcRERERERGxU062LkBur+3ILbYuQcQmYkL9bV2CiIiIiMhDzZBP6vft28fw4cOtts2ePZtNmzbd8ryBAwcycOBAq23+/v588MEHltfHjh0jKCgIgNDQUNq2bUtQUBDdu3dn8ODBJCUl3fIa/v7+9OjRg549exIQEMDq1avvZmoiIiIiIiIiFnpS/18nTpzg8uXLZGdnk5SUhK+vr2XfypUr+cc//kGlSpXynTd69GiaNGkCwHfffUdwcDAbN2685bViYmJwdXUlKyuLNm3a8NJLL1GqVKnCnZCIiIiIiIgYnkL9f23YsIHmzZvj5uZGXFwcISEhln2hoaGEhoayZs2aW47x7LPP4uzsTGJiIk888cRtr3n16lVcXV3x9PQscP0iRuTtrf/bkIeTelOMQH0sRqA+FiMoaB8bNtTv3bvXskweICkpibfeeuuGx+bl5bF9+3bWrVuHk5MTL7/8MsOGDcPNzQ2Apk2b8uWXX7Js2TJatmx5y+uWKlWKtLS0W4b6vn37YjKZ+P3332nRogXOzs73MEMR40tNvWTrEkTy8fb2VG+K3VMfixGoj8UIrvdxQYK9YUP9888/T1RUlOX17Nmzb3rsrl27yMjIYOTIkcC1kL9t2zY6d+5sOSY0NJSOHTtSoUKFW143JSUFHx+fWx7z5+X3AwYMYOvWrbRv3/5OpiUiIiIiIiJiYdhQfzc2bNhAeHg4L7zwAgD79+8nPDzcKtR7eHgwZcoURowYccPP1gPs3r0bNze324b661xcXChVqhTZ2dkFnoOIiIiIiIg8eh75UH/27Fl++OEHq6f6zzzzDJmZmRw4cMDq2AYNGvDyyy/zyy+/WLbNmjWLZcuW4eDggLu7O++8885tr9m3b18cHBzIy8vDx8eHdu3a3fL4bXPaa2mRGIKWyYmIiIiIFC6T2Ww227oIuT0FITEChXoxAvWxGIH6WIxAfSxGoM/U34WsrCz69euXb7ufnx9Tpkwp1Gt99tlnrFy5Mt/2Xr163faL9kRERERERETulJ7U2wm9CylGoHfUxQjUx2IE6mMxAvWxGEFhPKl3KMR6REREREREROQBUqgXERERERERsVMK9SIiIiIiIiJ26qH/orx9+/bRq1cvoqKiaNOmjWV727ZtqVmzJhEREZw6dYpWrVoRERFB69atLcckJSUxa9YsTp48iZubG25ubowePZonn3zScszAgQMBWLx4sWVb//79yczMBCAnJ4cDBw6QkJDAsmXLaNOmDZUqVeLFF19k3bp11KpVC4A1a9Zw5swZhg4dir+/P+XKlcPB4X/vmYSEhFCrVi2WLl3Knj17cHBwwGQyMXz4cMsYN9N25JYC3EER44oJ9bd1CSIiIiIiNvXQh3qASpUqsX37dkuo//XXX7ly5Ypl/6ZNm+jVqxdxcXGWUH/lyhUGDRrE1KlTqVu3LgAHDx5kypQpxMbGAnDixAkuX75MdnY2SUlJ+Pr6AvDee+9Zxp44cSI1atSgcuXKVjV5eHjw9ttvs3HjRlxcXPLVHBMTg6urq9W2o0ePsnPnTtasWYPJZOKXX34hJCSErVu3FvQWiYiIiIiIyCPILpbfV69enRMnTnDx4kUAtm7dStu2bQEwm81s2bKFPn36kJ2dzZEjRwD4/PPPef755y2BHqB27dp88MEHltcbNmygefPmdOjQgbi4uHzXjYuL4/fff+ftt9/Ot++JJ56gcePGREVF3fE8vLy8SElJYcOGDZw6dYoaNWqwYcOGOz5fRERERERE5M/sItQDtGzZkk8//RSz2czBgwctYf3rr7+matWqeHl50bFjR1avXg1AcnIyFSpUsJw/aNAggoKCeOmllzh58iR5eXls376d9u3b8/LLL5OQkMDVq1ctx+/fv58VK1Ywb948nJ2db1hTcHAwu3fv5rvvvsu3r2/fvgQFBREUFMRrr70GXAv10dHRHDhwgK5du/LSSy/x+eefF9o9EhERERERkUeLXSy/h2ufoQ8LC8PX15dnn33Wsn39+vUkJyfTr18/srOzOXz4MKNGjcLHx4effvrJclx0dDQAXbp0IScnh127dpGRkcHIkSMByMvLY9u2bXTu3JlTp04xZswY5s2bR6lSpW5ak4uLCzNmzGDkyJF06dLFat+Nlt8nJibi4eHBjBkzAPjxxx8ZMGAADRo0oESJEgW6PyKPooL8nqdIQaj3xAjUx2IE6mMxgoL2sd2Eel9fXy5fvkxsbCwjRowgKSmJtLQ0jhw5wo4dO3B0dARg/PjxbN68mY4dO7Js2TK+//576tSpA1wL1SdPnsRkMrFhwwbCw8N54YUXgGtP5sPDw2nfvj3Dhg1j2LBht/0CO4CaNWvyyiuvsGzZMgIDA2957K+//sqaNWtYvHgxrq6u+Pn54enpaaldRO5OauolW5cgjyBvb0/1ntg99bEYgfpYjOB6Hxck2NtNqAdo06YNW7Zswc/Pj6SkJL799ls6d+5sFYq7dOnCmDFjCAoKIjo6mjlz5jB79mxycnJwcnJi6tSpuLm58cMPP1h9Hv6ZZ54hMzOT5cuX8+uvvxIfH098fLxl/7Bhw25a18CBA/Mto+/bt6/Vt9/36tWLVq1acezYMTp37kzRokUxm82MGTMGT0+9wygiIiIiIiJ3z2Q2m822LkJuT+9CihHoHXUxAvWxGIH6WIxAfSxGUBhP6u3mi/JERERERERExJpCvYiIiIiIiIidUqgXERERERERsVMK9SIiIiIiIiJ2SqFeRERERERExE4p1IuIiIiIiIjYKbv6nfpHVduRW2xdgogUQEyov61LEBERERGDUqi/jX379tGrVy+ioqJo06aNZXvbtm2pWbMm33zzDeXKlcPB4X+LHkJCQqhVqxYAS5cu5YMPPuCzzz7D1dWVkSNHcvr0aY4fP46zszNlypShatWqTJgw4YHPTUREREREROybQv0dqFSpEtu3b7eE+l9//ZUrV65Y9sfExODq6nrDc7dt20abNm348MMPCQgIYM6cOQAsWLCA0qVL07179/s/ARERERERETEkfab+DlSvXp0TJ05w8eJFALZu3Urbtm1ve96+ffuoUKEC3bp1Y/Xq1fe7TBEREREREXnE6En9HWrZsiWffvopAQEBHDx4kNdff50TJ04A0LdvX8vyewcHB95//30A4uPj6dy5M5UqVcLFxYUffviBp59+2mZzEBHb8Pb2tHUJUsj0NxUjUB+LEaiPxQgK2scK9Xeobdu2hIWF4evry7PPPmu170bL7y9cuMCXX37JuXPniI2NJT09nVWrVinUizyCUlMv2boEKUTe3p76m4rdUx+LEaiPxQiu93FBgr1C/R3y9fXl8uXLxMbGMmLECJKSkm55/NatW+nYsSMhISEAXLlyhebNm3Pu3Dm8vLweRMkiIiIiIiJicAr1d6FNmzZs2bIFPz8/q1D/5+X3AL169SI+Pp6ZM2dathUpUoRWrVqxfv16Bg4c+EDrFhEREREREWMymc1ms62LkNvT0iIxAi2TEyNQH4sRqI/FCNTHYgSFsfxe334vIiIiIiIiYqcU6kVERERERETslEK9iIiIiIiIiJ1SqBcRERERERGxUwr1IiIiIiIiInZKoV5ERERERETETt3T79QnJSUxa9YsTp48iZubG25ubowePZonn3ySBQsWsH37dsqUKWM5vmHDhgwaNAh/f3969+5Nr169ADh27BhhYWHExsYSGhrKoUOHKFGihOW8du3a0blzZ2rVqkXdunUByM7OJi8vjzlz5uDr6wvAqVOnaNWqFREREbRu3RqAffv2sXbtWqKiom44h02bNvH222+zfv16nn76acvY//jHP+jZsydDhw61uu51s2fPZvfu3cyfP99y/YsXL1KvXj0mTZoEwG+//casWbO4cuUKly9fpmnTpgwdOpTjx48zYsQI1q9fbxlvzZo1nDlzhqFDh970frcdueX2fxQRsRsxof62LkFEREREDOKuQ/2VK1cYNGgQU6dOtQTegwcPMmXKFGJjYwHo3bs33bt3v+H5K1eu5B//+AeVKlXKt2/06NE0adIk3/bixYtbxgZYu3YtK1asYOLEicC1gN6rVy/i4uIsof5OVKpUie3bt1tC/a5du/D0/N/vA/71un/2yiuvMGrUKADy8vIIDAzkxx9/5IknnmDEiBEsWLCAihUrkpuby7Bhw1i7di2NGze+49pEREREREREbueul99//vnnPP/881ZPsGvXrs0HH3xwR+eHhoYSGhpKbm7u3V7aIiUlhWLFigFgNpvZsmULffr0ITs7myNHjtzxOE2aNGHPnj3k5eUB8OGHH/Lyyy/fdT0ZGRlcunQJT09PPvvsMxo0aEDFihUBcHR0JDIyko4dO971uCIiIiIiIiK3ctdP6pOTk6lQoYLl9aBBg0hPT+f06dO8//77wLWn8QkJCZZjBg4cSKNGjQBo2rQpX375JcuWLaNly5ZWY8+aNYtly5ZZXo8fP55q1apx4cIFgoKCSE9P5/z587Rq1Yq33noLgK+//pqqVavi5eVFx44dWb16NZMnT76juTg7O1OnTh2++eYbatWqRXp6Oj4+Ppw5cwbAct3rypQpw5w5cwDYvn0733//Pampqbi7uzNw4EAqVqzIv/71L8uy/Ovc3d0t/z569KjVmKdPn+aVV165o3pFRERERERE/uyuQ72Pjw8//fST5XV0dDQAXbp0IScnB7j18nu49rS+Y8eOVm8OwO2X3+fm5hIaGoqzs7MlKK9fv57k5GT69etHdnY2hw8ftiyLvxOvvPIKH374ISdOnKBly5ZkZ2fnu+7Nzhs1ahRJSUn079/f8mS+fPny/Pzzz1bHJiUlcfLkScqVK0eVKlWsxrz+mXoReXR4e3ve/iB5qOlvKEagPhYjUB+LERS0j+861Ddv3pxly5bx/fffU6dOHQASExM5efIkJpPpjsbw8PBgypQpjBgx4oafrb8ZR0dHpk6dSvv27Xn22WepXbs2P/zwAzt27MDR0RG49nR/8+bNVKtW7Y7GbNCgAdOnT+f06dPMmTOHbdu23XE9AL6+vkyaNIlhw4bx4Ycf0qxZM5YsWUL37t2pUKEC2dnZRERE0LBhQ8qVK3dXY4uIMaWmXrJ1CVIA3t6e+huK3VMfixGoj8UIrvdxQYL9XYd6d3d3oqOjmTNnDrNnzyYnJwcnJyemTp3KY489BuRffu/n58eUKVOsxmnQoAEvv/wyv/zyi2XbX5ffP/fcc5Zl9te5ubkxbdo0QkJC6NSpE61atbIEeri2YmDMmDGEhYWxe/duAgICLPvmzJmDn5+f1XgODg40atSIEydO4OHhYbXvr8vvAUaMGJHvnjRs2JCGDRsyf/58QkJCiIiIYPz48ZjNZjIyMmjWrBmBgYEcP378xjdVRERERERE5B6YzGaz2dZFyO3pXUgxAr2jLkagPhYjUB+LEaiPxQgK40n9XX/7vYiIiIiIiIg8HBTqRUREREREROyUQr2IiIiIiIiInVKoFxEREREREbFTCvUiIiIiIiIidkqhXkRERERERMRO3fXv1MuD13bkFluXICJ2IibU39YliIiIiMgDpFAP7Nu3j+DgYKpUqYLZbCYnJ4dp06axbNkyDh06RIkSJTCbzZw/f54+ffrQsWNHFixYwPbt2ylTpoxlnIYNGzJo0CAATp06RatWrYiIiKB169Y3vU6vXr1o06aNTeYtIiIiIiIi9k2h/r+ef/55oqKiAPjqq6+YOXMmJUuWZPTo0TRp0gSA8+fP88orrxAQEABA79696d69+w3H27RpE7169SIuLs4S6v96nYyMDIKCgvDz86NGjRr3c3oiIiIiIiJiQPpM/Q1cvHiRxx57LN/2M2fO4OLigslkuuX5ZrOZLVu20KdPH7Kzszly5MgNj3N3d6dr1658/PHHhVK3iIiIiIiIPFr0pP6/9u7dS1BQEFlZWfz6668sWbKEzZs3M2vWLBYvXkxKSgqVK1dm3rx5lnNWrlxJQkKC5fXAgQNp1KgRX3/9NVWrVsXLy4uOHTuyevVqJk+efMPrlipVikOHDt33+YnIo8Hb29PWJTwSdJ/FCNTHYgTqYzGCgvaxQv1//XlZ/O+//063bt1o2LChZfn9F198wezZs6lQoYLlnJstv1+/fj3Jycn069eP7OxsDh8+zKhRo2543ZSUFHx8fO7PpETkkZOaesnWJRiet7en7rPYPfWxGIH6WIzgeh8XJNhr+f0NlC5dOt+2pk2b0rx5cyZMmHDLc8+dO8cPP/xAfHw8y5cv54MPPqBVq1Zs3rw537Hp6enEx8fz0ksvFVrtIiIiIiIi8ujQk/r/ur783sHBgYyMDEJDQ/nmm2+sjhk8eDABAQH8+9//BvIvv/fz88PPz49WrVrh6Oho2d6lSxfGjBlDWFiY1XVyc3MZOnQolSpVeiBzFBEREREREWMxmc1ms62LkNvT0iIxAi2TEyNQH4sRqI/FCNTHYgRafi8iIiIiIiLyCFOoFxEREREREbFTCvUiIiIiIiIidkqhXkRERERERMROKdSLiIiIiIiI2CmFehERERERERE7pd+p/699+/axdu1aoqKiLNvMZjNxcXFs374dJ6drt6p///40bdqUf/7zn2zcuJHMzEyOHj1KzZo1AQgICGDTpk0A/N///R9169YFICQkhGrVqrFkyRL27NmDo6MjTk5OBAcH8/TTT9+ytrYjt9yPKYuIQcWE+tu6BBERERF5QBTqb2HdunUcOHCAlStX4urqSlpaGgMGDKB48eJ06NCBDh06kJyczIgRI4iNjbWc9+qrrwLQqFEjq+1z5swhNzeXVatW4eDgwPHjx3njjTeIjo7G19f3gc9PRERERERE7JuW39/CqlWrGDduHK6urgCULFmSIUOGsGbNmnsab+vWrYwYMQIHh2u3/bHHHiMwMJDNmzcXWs0iIiIiIiLy6FCov4W0tDS8vLystvn6+pKSknLXY509e5bixYtblvEXdDwRERERERERLb+/BQ8PD86fP0+JEiUs2xITEylXrtxdj+Xp6cmFCxfIycmxCvb3Op6IyM14e3vaugTD0z0WI1AfixGoj8UICtrHCvW30LNnT8LDw5k+fTouLi6cPXuWhQsXMm7cuLsey8XFhdatWxMVFcXIkSNxcHAgKSmJuLg4lixZch+qF5FHVWrqJVuXYGje3p66x2L31MdiBOpjMYLrfVyQYK9Q/ye7d+8mICDA8vr6F9v16NEDJycnTCYTgwcPpl69evc0/qhRo1iwYAFdunTB2dkZFxcXwsPD9SV5IiIiIiIick9MZrPZbOsi5Pb0LqQYgd5RFyNQH4sRqI/FCNTHYgSF8aReX5QnIiIiIiIiYqcU6kVERERERETslEK9iIiIiIiIiJ1SqBcRERERERGxUwr1IiIiIiIiInZKoV5ERERERETETul36u1A25FbbF2CiEihign1t3UJIiIiIobwyIT65ORk2rVrR82aNS3bGjRoQExMjGVbZmYmRYsWZd68eRQvXpyMjAyioqL45ZdfcHBwwN3dnZCQEPz8/Pjmm29YsGCBZayTJ09SokQJoqKi8l0HYOXKlSxatIgvvviCtWvX4uR07dZ36dKFuXPn8vjjjz+AuyAiIiIiIiJG8siEeoAqVaoQGxtreZ2cnMyXX35ptW3OnDls2LCBfv36MWHCBOrWrcv48eMBOHz4MG+++Sbr1q2jfv36lvPOnDlDYGAgoaGhN7zOnx0/fpwlS5bw5ptv3q9pioiIiIiIyCNCn6n/E7PZzIkTJyhWrBjnzp3jyJEjBAUFWfZXr16dZs2a8cknn1i2ZWdn89Zbb9GvXz+eeeaZ216jf//+bNu2jZ9//vm+zEFEREREREQeHY/Uk/qjR49ahfTg4GDLtvPnz5OZmUnbtm159dVX+fnnn/H19c03hq+vLykpKZbX06ZNo0qVKnTt2vWm16lZs6blKX7RokUJDw8nNDSUDRs23I9piog89Ly9PW1dQoHYe/0ioD4WY1AfixEUtI8fqVB/o+X317ddvXqVgQMHUqpUKZycnChTpoxVeL8uMTGRypUrA7Bx40Z+/fVXPvjgg1te56+effZZGjZsyLx58wppZiIi9iU19ZKtS7hn3t6edl2/CKiPxRjUx2IE1/u4IMFey+//y83NjdmzZ7No0SIOHz6Mj48PFSpUYPXq1ZZjDh06xM6dO2nVqhUHDx5kyZIlLFiwAGdn57u+3vDhw/nyyy9JTEwszGmIiIiIiIjII+SRelJ/O6VLl2bMmDFMnDiRtWvXEhkZycyZM+ncuTOOjo4UK1aMRYsWUaxYMaKiojCbzQwfPtxyftGiRZkwYUK+5fcA06dPt3rt6urK9OnT6dat2wOZm4iIiIiIiBiPyWw2m21dhNyelhaJEWiZnBiB+liMQH0sRqA+FiPQ8nsRERERERGRR5hCvYiIiIiIiIidUqgXERERERERsVMK9SIiIiIiIiJ2SqFeRERERERExE4p1IuIiIiIiIjYKf1O/Q3s27eP4OBgqlSpgtlsJicnh169etGmTRsAEhISGDt2LP/6178oW7YsAJs2beL3339n1KhR+Pv7U65cORwcHMjNzeXy5ctMnTqV1NRUoqKi2LhxIy4uLgDMmDEDZ2dnRo0addN62o7ccv8nLSJiUDGh/rYuQUREROS+Uai/ieeff56oqCgAMjIyCAoKws/Pjxo1ahAfH0/Pnj1Zv349Q4cOveH5MTExuLq6ArBr1y4WLlzIkiVL2LFjB4sWLSI4OJgDBw6wf/9+1qxZ88DmJSIiIiIiIsah5fd3wN3dna5du/Lxxx+TlJTEhQsXeOONN9iyZQvZ2dm3PT8lJYVixYoBMHbsWBISEjh06BDh4eFERkbi7Ox8v6cgIiIiIiIiBqRQf4dKlSpFWloaGzZsoGPHjnh6elKnTh0+/fTTGx7ft29fOnXqRJMmTTh48CAhISEAeHh4EB4ezmuvvUbnzp2pXLnyg5yGiIiIiIiIGIiW39+hlJQUypQpw6ZNm3jsscfYuXMnFy5cYNWqVZbP2v/Z9eX3c+fOJTk5mVKlSln21a9fn2LFihEQEPAgpyAi8kjy9va0izFFHjT1sRiB+liMoKB9rFB/B9LT04mPj6dTp07UqlWL+fPnW/a9+OKLHD58+KbnBgcH06tXL+Li4ujRo8eDKFdERP4kNfVSoY7n7e1Z6GOKPGjqYzEC9bEYwfU+LkiwV6i/ib179xIUFGT5BvuhQ4eyadMmOnfubHVcp06dWL16NXXr1r3hOA4ODkybNo0ePXrQokULy7fli4iIiIiIiBSUyWw2m21dhNye3oUUI9A76mIE6mMxAvWxGIH6WIygMJ7U64vyREREREREROyUQr2IiIiIiIiInVKoFxEREREREbFTCvUiIiIiIiIidkqhXkRERERERMROKdSLiIiIiIiI2Cn9Tr0daDtyi61LEBGRuxQT6m/rEkREROQR8FCE+qVLl7Jnzx4cHBwwmUwMHz6cWrVqAfDRRx+xatUqHBwcyMnJoWvXrnTo0AGAoKAgrly5QpEiRSxjjRs3jmnTpgHw/fffU7t2bRwcHOjXrx8vvPCC5bhatWpRt25dAK5evco//vEPhg4dioODA/7+/pQrVw4HBwdyc3O5fPkyU6dO5W9/+xuhoaG0adOGSpUq8eKLL7Ju3TpLrWvWrOHMmTMMHToUf39/PvroI1xdXUlNTaVv377079+f9u3b33K+IiIiIiIiInfK5qH+6NGj7Ny5kzVr1mAymfjll18ICQlh69atfPXVV6xdu5bFixfj6enJ1atXeeutt3B1daV169YAREZGUrlyZasxY2NjAfD39ycmJgZXV9d81y1evLjlOLPZzKRJk1i9ejVBQUEAVuft2rWLhQsXsmTJEqsxPDw8ePvtt9m4cSMuLi43nN+pU6d4/fXXGTZsGC1atLjlfEVERERERETuhs0/U+/l5UVKSgobNmzg1KlT1KhRgw0bNgDXwvmoUaPw9PQEwM3NjZCQEFavXl2oNZhMJvr06UNCQsIN96ekpFCsWLF825944gkaN25MVFTUTc/r06cPoaGhtGjRArj1fEVERERERETuhs2f1Ht5eREdHc2qVat49913cXNzY/jw4bz44oskJSVRoUIFq+N9fX1JSUmxvA4JCbEsv2/Xrh2dO3e+pzpKly5NWlqa5XXfvn3JzMzk9OnTNG7cmJCQkBueFxwcTKdOnfjuu+/y7Xvrrbdwc3Pj7NmzdzRfERExDm9vT1uXIHJL6lExAvWxGEFB+9jmoT4xMREPDw9mzJgBwI8//siAAQNo0KABZcuW5fjx4xQvXtxy/B9//EG5cuUsr2+0/P5eHD9+HB8fH8vr68vv586dS3JyMqVKlbrheS4uLsyYMYORI0fSpUsXq33Tp0+ndOnSdO/enaeeeorKlSvfcr4lSpQo8DxEROThkJp6ydYliNyUt7enelTsnvpYjOB6Hxck2Nt8+f2vv/5KWFgYmZmZAPj5+eHp6YmjoyNBQUHMnDmT9PR0ADIyMpg5cyY9evQo1Bry8vKIiYnh5ZdfzrcvODiY06dPExcXd9Pza9asySuvvMKyZcustletWpVy5coRGhpKcHAwV69eveV8RURERERERO6GzZ/Ut2rVimPHjtG5c2eKFi2K2WxmzJgxeHp64u/vT3p6Ov3798dkMpGXl0enTp1o06ZNga974cIFgoKCMJlM5OTk0LBhQzp16pTvOAcHB6ZNm0aPHj0sn4u/kYEDB/L555/fcN9LL73Erl27mDx5MjNmzLjpfEVERERERETuhslsNpttXYTcnpYWiRFomZwYgfpYjEB9LEagPhYjMMTyexERERERERG5Nwr1IiIiIiIiInZKoV5ERERERETETinUi4iIiIiIiNgphXoRERERERERO6VQLyIiIiIiImKn7up36pOTkxkxYgTr168nNDSU9PR0Fi5caNnfqFEjdu/eTV5eHpGRkRw5cgQHBwecnZ0ZN24cly9fJjw8HIDvv/+e2rVr4+DgQL9+/WjatClxcXFs374dJ6drZfXv35+mTZsCUKtWLerWrQtAdnY2eXl5zJkzB19fX/z9/SlXrhwODv97jyIkJIRatWpZXi9YsIDt27dTpkwZcnJyKFWqFBEREXh4eABw6tQpWrVqRUREBK1btwZg3759rF27lqioKAA+/vhjFi5cyNKlS5k/fz6HDh2iRIkSlmu0a9eOzp0737TWf//733zxxRe89957lnOGDh1Kw4YN6d69+03ve9uRW+7mzyQiInYmJtTf1iWIiIiInbqrUP9X+/fv55///CcdOnSw2r5r1y5Onz7NihUrANixYwfTp08nOjqa2NhYAPz9/YmJicHV1RWAtWvXcuDAAVauXImrqytpaWkMGDCA4sWLU6dOHYoXL2459/rxK1asYOLEiQBWY91M7969LeF57ty5rFu3jn79+gGwadMmevXqRVxcnCXU/9mHH37I8uXLWblyJaVLlwZg9OjRNGnSJN+xN6t1woQJ7Ny5k/j4eDp37syHH35Idnb2LQO9iIiIiIiIyM0UaPn9yJEjWbBgASdPnrTa7uPjw08//URCQgLnzp2jefPmzJs375ZjrVq1inHjxlmCecmSJRkyZAhr1qy54fEpKSkUK1bsnmu/cOECjz32GABms5ktW7bQp08fsrOzOXLkiNWx//znP1mxYgUrVqywBPq7cb1Wk8nEjBkziI6O5ujRoyxevJjp06ff8xxERERERETk0VagJ/VlypRh2LBhjBs3juXLl1u2V6tWjalTp7J+/XrCw8Px8fEhNDSU+vXr33SstLQ0vLy8rLb5+vqSkpICXAvhQUFBpKenc/78eVq1asVbb71lObZv376W5fcODg68//77+a6xcuVKEhISOH/+PJcvX2bw4MEAfP3111StWhUvLy86duzI6tWrmTx5MgDfffcdp06d4sKFC+Tm5lqNN2vWLJYtW2Z5PX78eKpVq3bLWn18fHjrrbfo2rUrc+fOzTdnERERERERkTtVoFAP1z5HvmPHDuLi4izbDh8+jJ+fH3PnzsVsNrN7926Cg4PZvXs3JpPphuN4eHhw/vx5q8+oJyYmUq5cOeB/S9pzc3MJDQ3F2dkZd3d3y7F3u/w+Pj6ekJAQVq5cyfr160lOTqZfv35kZ2dz+PBhRo0aBYC3tzcrVqwgPj6e0aNHs2zZMsubB7dbfn+zWjt06MCsWbMs3xcgIiKPNm9vT1uXII8g9Z0YgfpYjKCgfVzgUA8QFhZGly5dyMjIAK49+T58+DDTp0/H0dGRJ598kiJFitw00AP07NmT8PBwpk+fjouLC2fPnmXhwoWMGzfO6jhHR0emTp1K+/btefbZZ3nhhRfuqeby5cuTnZ3NuXPn+OGHH9ixYweOjo7AtSfumzdvplq1ajzxxBO4urrSs2dPvvrqK6Kjo3nzzTfv6BqFVauIiBhbauolW5cgjxhvb0/1ndg99bEYwfU+LkiwL5RQ7+XlRWhoqCXsBgUFERkZSYcOHfDw8MDBwYGZM2fecoygoCByc3Pp0aMHTk5OmEwmBg8eTL169fId6+bmxrRp0wgJCbEs6f/z8nuAXr160bJlS6vzri+/d3R05OrVq4wdO5YtW7bQqlUrS6AH6NKlC2PGjCEsLMzq/OnTp9OhQweeeeYZIP/y++eee87qIwE3qrVo0aK3vA8iIiIiIiIid8pkNpvNti5Cbk/vQooR6B11MQL1sRiB+liMQH0sRlAYT+oL9O33IiIiIiIiImI7CvUiIiIiIiIidkqhXkRERERERMROKdSLiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcK5Xfq5f5qO3KLrUsQEZEHJCbU39YliIiIiB2xWahfunQpe/bswcHBAZPJxPDhw1m1ahXp6eksXLjQclyjRo3YvXu35fWkSZP44Ycf+Oc//2nZlpeXx9KlS/nyyy9xdHQEYPz48VSrVo3Q0FDatGlDkyZNyMnJYeTIkZQsWZJJkybRvHlzypUrh4PDtQULxYsXZ+HChWRnZ7NkyRL27NmDo6MjTk5OBAcH8/TTTwNw7tw5IiMjSUlJITc3l3LlyhEaGoq3tzebNm1i/vz5+Pr6kpeXh8lk4s033+Tvf//7Teddq1at+327RURERERExIBsEuqPHj3Kzp07WbNmDSaTiV9++YWQkBCeeuop9u/fzz//+U86dOiQ77wrV65w4MABqlatyr59+2jQoAEA7733HmlpaaxatQoHBwcOHjzI4MGD+fjjjy3nZmdnM3z4cCpWrMioUaMs22NiYnB1dbW6zvz588nNzbWMd/z4cd544w2io6N5/PHHGTJkCH379qVFixYA7NmzhzfeeIP4+HgAXnnlFcs1zpw5Q48ePVi1ahUXLly44by3bt1aqPdXREREREREHg02+Uy9l5cXKSkpbNiwgVOnTlGjRg02bNgAwMiRI1mwYAEnT57Md95HH33E3//+d1599VVWr15t2b5u3ToGDx5seeJeu3ZtNmzYgLOzMwBZWVkMHTqU6tWrWwX6m9m6dSsjRoywjPfYY48RGBjI5s2b+emnn/D09LQEeoCGDRtSoUIFvv3223xjlS5dmhdffJF///vft5y3iIiIiIiIyN2yyZN6Ly8voqOjWbVqFe+++y5ubm4MHz4cgDJlyjBs2DDGjRvH8uXLrc6Lj49nypQpVK5cmbCwME6dOkXZsmW5evUqxYsXtzq2ZMmSln9PmzYNX19fTp06la+Wvn37WsJ7v379+Nvf/kbx4sVxcrK+Nb6+vhw8eJCkpCR8fX3zjePr60tKSsoN51uqVCnS0tJuOu8XX3zxDu6aiIg8Cry9PW1dgjwi1GtiBOpjMYKC9rFNQn1iYiIeHh7MmDEDgB9//JEBAwZYPrPerl07duzYQVxcnOWcY8eO8dtvvxEREQGAyWRizZo1BAcHU6xYMdLT0/Hw8LAc/+mnn1o+x96zZ09ee+01evTowZYtW2jfvr3luL8uv8/KyuLChQvk5ORYBfvExETKlStH2bJlOX78+A3n1LBhQ06cOJFvX0pKCk899dRN592gQQNKlChx1/dRRESMJzX1kq1LkEeAt7enek3snvpYjOB6Hxck2Ntk+f2vv/5KWFgYmZmZAPj5+eHp6Wn5kjuAsLAwYmJiyMjIAK49pR8+fDjLly9n+fLlvP/++2zcuJGsrCxeffVVFi5ciNlsBuDAgQPMmDEDFxcXAJ588kmcnJyYPXs2s2bN4tixYzetzcXFhdatWxMVFUVeXh4ASUlJxMXFERAQQL169Thz5gw7d+60nPPll1+SmJhI/fr18413+vRpPvvsM5o2bXpH8xYRERERERG5UzZ5Ut+qVSuOHTtG586dKVq0KGazmTFjxrBjxw7LMV5eXoSGhvLmm2+SlZXFhx9+yJYt//tpt/Lly1O9enX+9a9/0a9fP+bNm0fXrl1xcnLCycmJ6OhoS6i/ztfXl9GjRzNs2DDLl9rdyKhRo1iwYAFdunTB2dkZFxcXwsPDLcvuFy9ezPTp01myZAkAPj4+LF261BLOt2/fzg8//ICDgwNms5kZM2ZQokSJm87b01PLhkREREREROTumczXH2/LQ01Li8QItExOjEB9LEagPhYjUB+LEdjt8nsRERERERERKTiFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETslEK9iIiIiIiIiJ1SqBcRERERERGxUzb5nXp7t2/fPoKDg6lSpQpms5mcnBx69epF7dq1adeuHTVr1rQ6fuXKlSxatIjSpUvTvXt3q33+/v589NFHuLq63vR6bUduuS/zEBGRh1dMqL+tSxARERE7oFB/j55//nmioqIAyMjIICgoiGnTplGlShViY2NtXJ2IiIiIiIg8CrT8vhC4u7vTtWtXli9fbutSRERERERE5BGiJ/WFpFSpUqSlpXH06FGCgoIs22vWrEloaKgNKxMRERERERGjUqgvJCkpKTzzzDNcunRJy+9FRKTAvL09bV2CGJx6TIxAfSxGUNA+VqgvBOnp6cTHxzNv3jz+/e9/27ocERExgNTUS7YuQQzM29tTPSZ2T30sRnC9jwsS7BXq79HevXsJCgrCwcGB3Nxchg4diouLS77l9wDTp08HYOnSpcTHxwPXPoevJ/oiIiIiIiJSECaz2Wy2dRFye3oXUoxA76iLEaiPxQjUx2IE6mMxgsJ4Uq9vvxcRERERERGxUwr1IiIiIiIiInZKoV5ERERERETETinUi4iIiIiIiNgphXoRERERERERO6VQLyIiIiIiImKn9Dv1dqDtyC22LkFERB5CMaH+ti5BREREbExP6kVERERERETslCFDfUpKCjt37rzj47t06UJycnKBrvntt99y+PBhAIYMGVKgsURERERERETuhCFD/d69ezlw4MADvebGjRs5ffo0AAsXLnyg1xYREREREZFH00P7mfqrV6/y9ttvk5KSQnZ2Ni+++CKXLl1i1KhRZGZm0rp1a3bu3Mnq1av55z//iYODA/Xq1WPUqFEsXbqUq1evUrduXZo3b37D8aOioti1axc+Pj6kpaUBsGDBAkqXLk337t05duwYYWFhxMbG8sorr1CxYkVcXFwYM2YMYWFhZGZmcv78ed588018fHzYtWsXhw4dokqVKnTu3Jndu3fz888/M3XqVBwdHXF1dWXq1Knk5eUxcuRIfHx8SEpK4m9/+xuTJ09+kLdWREQMwtvb09YliB1T/4gRqI/FCAraxw9tqF+7di2PPfYYUVFRHDlyhD179nDp0qV8x23atIkJEyZQp04d4uLiMJvNDBgwgN9///2mgf7IkSN8++23bNiwgcuXL9OqVatb1nL58mUGDx7MU089xZ49e+jTpw8NGjTgwIEDLFiwgBUrVtC4cWPatGlD+fLlLeeNHz+eadOmUaNGDXbs2EFERARjxozhjz/+YPny5RQpUoQWLVqQmpqKt7d3wW6YiIg8clJT8///RZE74e3tqf4Ru6c+FiO43scFCfYP7fL733//nTp16gBQtWpVihUrZtlnNpst/54xYwZr166lZ8+epKSkWO27maNHj1KrVi0cHBzw8PCgatWqtz3Hz88PAG9vb9atW8fo0aNZu3YtOTk5Nz3n9OnT1KhRA4DnnnuO3377DYAKFSrg4eGBo6Mj3t7eZGZm3vb6IiIiIiIiIn/10Ib6ypUr8+OPPwKQlJTE2LFjSU1NBeDQoUOW49avX8/kyZNZtWoVv/zyC//3f/+Hg4MDeXl5Nx3bz8+PgwcPkpeXx+XLlzl69CgArq6uN7wGgIPDtVs1b9482rdvz6xZs2jQoIHlTQSTyZTvDYUyZcpYvjzv22+/pWLFipZjRURERERERArqoV1+361bN8aOHUvPnj3Jzc0lPj6eyMhIunfvTs2aNXF3dwegWrVqdOrUiZIlS1K2bFmefvppPDw8iI6OpmbNmrz88sv5xq5RowYvvfQSnTp1okyZMpQqVQqA1q1bExwczLfffkutWrVuWNdLL73EtGnTWLJkCeXKlbN8Hv/pp59m9uzZPP7445Zjw8PDmTp1KmazGUdHR6ZPn35P92LbnPZaWiSGoGVyYgTqYxEREXmYmMx3sl5dbE7/A1KMQGFIjEB9LEagPhYjUB+LERTGZ+of2if1hWHdunVs37493/YRI0ZQt25dG1QkIiIiIiIiUngMHeq7du1K165dbV2GiIiIiIiIyH3x0H5RnoiIiIiIiIjcmkK9iIiIiIiIiJ1SqBcRERERERGxU4b5TH1SUhKzZs3i5MmTuLm54ebmxujRo/n444/Zvn07ZcqUAeD8+fO0adOGQYMGAXD48GFmz55NZmYm2dnZNGjQgDfffBMXFxfg2pftbd26FQcHB7Kzsxk+fDgNGjRgwYIFVuMCNGzYkEGDBuHv70+5cuVwcHAgNzeXy5cvM3XqVFJTU4mKimLjxo2W8WfMmIGzszOjRo266dzajtxyv26biIgYSEyov61LEBERkQfMEKH+ypUrDBo0iKlTp1q+1f7gwYNMmTKF+vXr07t3b7p37w5AVlYWbdq0oUuXLpjNZkaMGMG7776Ln58fZrOZd999lxkzZjBp0iQ+/PBDdu/ezcqVK3F2diYpKYmePXuyefNmAKtx/yomJgZXV1cAdu3axcKFC1myZAk7duxg0aJFBAcHc+DAAfbv38+aNWsewF0SERERERERozHE8vvPP/+c559/3upn6mrXrs0HH3yQ79i0tDRycnJwdXVly5YtdOzYET8/PwBMJhNvvvkmX3zxBVevXmXt2rUMHDgQZ2dnAHx9ffnnP/+Jl5fXXdWXkpJCsWLFABg7diwJCQkcOnSI8PBwIiMjLeOLiIiIiIiI3A1DPKlPTk6mQoUKlteDBg0iPT2d06dP8+yzz7J9+3Y+/PBDTpw4QdmyZQkPD8fDw4OkpCQaNWpkNZbJZMLb25szZ85w+vRpfH19rfaXLFnS8u+VK1eSkJBgeT1w4EDLeH379iUzM5PTp0/TuHFjQkJCAPDw8CA8PJzXXnuNkSNHUrly5UK/HyIiIiIiIvJoMESo9/Hx4aeffrK8jo6OBqBLly7k5uZalsn/9NNPjBgxgooVKwJQtmxZjh8/bjVWbm4up0+fpnTp0jz22GOcOHECT09Py/6vvvqKatWqAXe2/H7u3LkkJydTqlQpy7769etTrFgxAgICCmX+IiIiAN7enrc/SOS/1C9iBOpjMYKC9rEhQn3z5s1ZtmwZ33//PXXq1AEgMTGRkydPWj0Jr1WrFq+//jojRoxg7dq1vPrqq/Tp04cXXniBihUrYjabWbhwIU2aNMHNzY2OHTuyaNEiZs+ejZOTE//v//0/xo0bx6ZNm+64tuDgYHr16kVcXBw9evQo7KmLiIhYpKZesnUJYie8vT3VL2L31MdiBNf7uCDB3hCh3t3dnejoaObMmcPs2bPJycnBycmJqVOncvDgQatjO3fuzEcffcSaNWvo0aMHM2fOZPLkyVy9epXs7Gzq16/PuHHjAHj55ZdJTU0lMDAQZ2dncnNzmTVrluWp+1+X3/v5+TFlyhSr6zk4ODBt2jR69OhBixYtKFu27H2+GyIiIiIiIvKoMJnNZrOti5Db07uQYgR6R12MQH0sRqA+FiNQH4sRFMaTekN8+72IiIiIiIjIo0ihXkRERERERMROKdSLiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETslCF+p97o2o7cYusSRETEDsWE+tu6BBEREbnPHoon9fv27aNatWokJCRYbW/bti2hoaEABAUFcezYMfbt28fw4cPzjXF9P0BGRgY9e/Zk6dKld3T95ORk6tWrR1BQkOU/CxcuzLe9S5cu9O7dmwsXLgBQq1Ytq3OCgoI4deoUeXl5zJgxgz59+tCvXz8GDhxIUlJSvjoBMjMz8ffX/+gSERERERGRu/fQPKmvVKkS27dvp02bNgD8+uuvXLly5a7HSU9P5/XXX+eVV16hR48ed3xelSpViI2NtdqWnJycb/ucOXPYsGED/fr1o3jx4vnOAfjiiy84ffo0K1asAGDHjh1Mnz6d6Ojou56PiIiIiIiIyM08FE/qAapXr86JEye4ePEiAFu3bqVt27Z3NcalS5fo06cPXbp0uatAf6fMZjMnTpygWLFitzzOx8eHn376iYSEBM6dO0fz5s2ZN29eodcjIiIiIiIij7aH5kk9QMuWLfn0008JCAjg4MGDvP7665w4ceKOzx89ejSlS5fm1KlTd33to0ePEhQUZHk9e/Zsq+3nz58nMzOTtm3b8uqrrwJw4cIFq3PKlCnDnDlzqFatGlOnTmX9+vWEh4fj4+NDaGgo9evXByAkJIQiRYoAkJeXd9e1ioiI3Alvb09blyAPMfWHGIH6WIygoH38UIX6tm3bEhYWhq+vL88+++xdnz9y5Ej+8Y9/0LFjR+rVq2cJ0deNGzeO//znP5QsWZL58+db7bvd8vurV68ycOBASpUqhZPTtdt2s+X3hw8fxs/Pj7lz52I2m9m9ezfBwcHs3r0bgMjISCpXrgxc+0x969at73quIiIit5OaesnWJchDytvbU/0hdk99LEZwvY8LEuwfmuX3AL6+vly+fJnY2FjatWt31+c/+eSTeHh4EBkZyZgxYzh79qzV/mnTphEbG5sv0N8JNzc3Zs+ezaJFizh8+PAtj/3666+ZO3cuubm5mEwmnnzySYoUKYLJZLrr64qIiIiIiIjczEP1pB6gTZs2bNmyBT8/P8s3xv/V7t27CQgIsLyeM2eO1f46derQpUsXRo4cSUxMDA4OhfPeRenSpRkzZgwTJ05k7dq1+ZbfA4wYMYKgoCAiIyPp0KEDHh4eODg4MHPmzEKpQUREREREROQ6k9lsNtu6CLk9LS0SI9AyOTEC9bEYgfpYjEB9LEZguOX3IiIiIiIiInLnFOpFRERERERE7JRCvYiIiIiIiIidUqgXERERERERsVMK9SIiIiIiIiJ2SqFeRERERERExE4V6Hfq9+3bR3BwMFWqVAEgMzOTtm3bEhQURGhoKIcOHaJEiRKW49u1a4ezszMbN24kMzOTo0ePUrNmTQBmz55N2bJlAUhOTqZdu3aWfQANGjRgyJAhnDt3jsjISFJSUsjNzaVcuXKEhobi7e0NwOHDh5k9ezaZmZlkZ2fToEED3nzzTVxcXKxqysnJoWTJkrz99tv4+vpazSsvL4+xY8dy5MgRypQpw/z583Fxcck3/1OnTtGqVSsiIiJo3bq1Zftvv/3GrFmzuHLlCpcvX6Zp06YMHToUk8nEiRMniIiI4Ny5c1y9epWaNWsyduzYG45/XduRW+7yLyMiImItJtTf1iWIiIjIfVCgUA/w/PPPExUVBUBWVhYvvfQS7du3B2D06NE0adIk3zkdOnQgOTmZESNGEBsbe8Nxq1Spkm+f2WxmyJAh9O3blxYtWgCwZ88e3njjDeLj40lLS2PEiBG8++67+Pn5YTabeffdd5kxYwaTJk3KV9N3331HcHAwGzdutLrO//t//4/ffvuNTZs2ERQUxJEjR6hVq1a+Gjdt2kSvXr2Ii4uzhPqLFy8yYsQIFixYQMWKFcnNzWXYsGGsXbuWLl26MHjwYMLCwnj66acBCA8PZ/78+YwaNerObriIiIiIiIjIfxU41P9Zeno6Dg4OODo6FuawFj/99BOenp6WQA/QsGFDKlSowLfffsuhQ4fo2LEjfn5+AJhMJt58802aN2/O1atX84337LPP4uzsTGJiIk888YRl+xNPPIHZbGby5MkUL16cGjVq5DvXbDazZcsW4uLiGDx4MEeOHKFq1ap89tlnNGjQgIoVKwLg6OhIZGQkzs7O7N+/Hx8fH0ugh2tvMuTl5RXWLRIREREREZFHSIFD/d69ewkKCsJkMuHs7MyECRNwd3cHYNasWSxbtsxy7Pjx46lWrdodjXv06FGCgoIsr2fPnk1SUlK+pfIAvr6+pKSkkJSURKNGjaz2mUwmvL29OXPmzA2vU6pUKdLS0qxC/b59+yhSpAg7d+5kzpw5nDx5kmLFiuHp6Wk55uuvv6Zq1ap4eXnRsWNHVq9ezeTJkzl9+nS+Gq/fjxvtc3V1vaP7ISIiIiIiIvJXhbr8/q9utvz+Ttxo+X3ZsmU5fvx4vmMTExNp2LAhp06dyrc/NzeX06dPU7p06RteJyUlBR8fH8vrq1evMmbMGLZt28aJEycYOnQoHh4eTJkyhTp16liOW79+PcnJyfTr14/s7GwOHz7MqFGjKF++PD///LPVNZKSkjh58iTly5fnk08+sdqXlpbG999/T7Nmze7ovoiIiNwLb2/P2x8kjxT1hBiB+liMoKB9XKjL7++3evXqcebMGXbu3Im//7Uv/Pnyyy9JTEykfv36+Pn50adPH1544QUqVqyI2Wxm4cKFNGnSBDc3t3zj7d69Gzc3N6tQn5eXh9lsxsHBgZo1a9KwYUM++ugjypQpYznm3Llz/PDDD+zYscPyUYPx48ezefNmAgICWLJkCd27d6dChQpkZ2cTERFBw4YN6d69O8nJyRw8eJDatWtb6nN1dVWoFxGR+yo19ZKtS5CHiLe3p3pC7J76WIzgeh8XJNjf11D/1+X3zz33HG+99dY9j2cymVi8eDHTp09nyZIlAPj4+LB06VIcHR3x8fFh5syZTJ48matXr5KdnU39+vUZN25cvpocHBxwd3fnnXfesbpG0aJFGT9+PP369cPZ2ZmqVasyevRo5syZw5w5cwDYsmULrVq1svrugC5dujBmzBiCgoKIiIhg/PjxmM1mMjIyaNasGYGBgZhMJubNm8eUKVMs34xfp04dgoOD7/meiIiIiIiIyKPLZDabzbYuQm5P70KKEegddTEC9bEYgfpYjEB9LEZQGE/qHQqxHhERERERERF5gBTqRUREREREROyUQr2IiIiIiIiInVKoFxEREREREbFTCvUiIiIiIiIidkqhXkRERERERMRO3dffqZfC0XbkFluXICIiIrcRE+pv6xJEROQR9NCF+uTkZNq1a0fNmjUt2xo0aECHDh2stmdmZlK0aFHmzZtH8eLFadSoEbt377ac8+WXX5KQkEBERARBQUGEhYURHx/PoUOHSE1N5erVq/j6+lKyZEnmz59vVcOqVavYsGEDjo6OvPPOO/j6+lrtHzNmDPXr16dTp06WbStXriQtLQ0nJye2b99OmTJlLPsaNmzIoEGDyMnJYfHixXzxxRe4uroC0LZtW7p27Vp4N1BEREREREQeGQ9dqAeoUqUKsbGxVtuSk5PzbZ8zZw4bNmygX79+dzRuaGgoAJs2beL3339n1KhRNzxu8eLFfPzxx6xYsYJPP/2Uvn37Wu3v0qUL8+bNswr1mzdv5t1332Xz5s307t2b7t275xs3KiqKvLw81q5di6OjIxkZGbzxxhs8++yzVK5c+Y7mICIiIiIiInLdQxnq74TZbObEiRNUqFCh0Md+6aWXGD9+PGfPnmXWrFn59j/77LOcO3eO48eP89hjj3Hw4EFKly7N448/ftMxc3Jy+Oijj/jkk09wdHQEwN3dndjYWEwmU6HPQURERERERIzvoQz1R48eJSgoyPJ69uzZVtvPnz9PZmYmbdu25dVXX73pOPcSlpOTkzl27Bhnz57l5Zdfxs3NjZMnT+Lj42N1XKdOndi6dSuDBg1i06ZNdOvWzbJv5cqVJCQkWF4PHDiQqlWrUrx4cZycrt3yuLg4PvroIzIyMmjXrh29e/e+61pFRETk4eHt7WnrEu6KvdUrciPqYzGCgvbxQxnqb7f8/urVqwwcOJBSpUpZQvJfA/zly5ctn1u/G+PHj6dPnz7UrVuXwMBAvvrqKxo3bsyAAQOsjmvfvj29e/emb9++fPPNN4wfP96y70bL77Ozszl//jy5ubk4OjoSGBhIYGAga9as4cyZM3ddp4iIiDxcUlMv2bqEO+bt7WlX9YrciPpYjOB6Hxck2NvlT9q5ubkxe/ZsFi1axOHDhwF4/PHH+frrry3H7Nq1i7/97W93PXZWVhbOzs4UK1aMAQMG8M0331CpUqV8x3l5eVG5cmUWLVpEy5YtLW8u3IyzszOtWrXinXfeIS8vD7j2ZX8//PCDlt+LiIiIiIjIPXkon9TfidKlSzNmzBgmTpzI2rVrCQ8PZ/LkyZYvo6tTpw7t27e/63EnTpzI1KlTWbhwIR4eHrz77rtER0fTpEkTXFxcrI7t0qULr7/+Oh9//LHV9r8uv/fz82PKlCmMHj2a9957jx49euDk5ER6ejotWrSgT58+93YTRERERERE5JFmMpvNZlsXIbenpUViBFomJ0agPhYjUB+LEaiPxQge2eX3IiIiIiIiIqJQLyIiIiIiImK3FOpFRERERERE7JRCvYiIiIiIiIidUqgXERERERERsVMK9SIiIiIiIiJ26p5/p/63335j1qxZXLlyhcuXL9O0aVOGDh1KWloakZGRpKSkkJubS7ly5QgNDcXb25tNmzaxcOFCtm7dioeHBwDDhw+nW7duNGjQgFq1alG3bl3LNSpXrkxYWFi+7QCzZ8+mbNmyfPfdd7z77rvk5ORw+fJlAgIC6NGjBxERERw6dIjU1FSuXr2Kr68vJUuWZP78+TRq1Ijdu3cDsHfvXhYtWoTZbCY7O5sXX3yR3r17YzKZCAoK4qmnnuLtt98GIDMzk9atW7Nz504WLFjA9u3bKVOmjKWmhg0bMmjQIPz9/SlXrhwODg7k5uZy+fJlpk6dSmpqKlFRUWzcuNHym/czZszA2dmZUaNG3fRetx255V7/TCIiImIAMaH+ti5BREQeUvcU6i9evMiIESNYsGABFStWJDc3l2HDhrFmzRq2b99O3759adGiBQB79uzhjTfeID4+HoArV64wffp0pk+fnm/c4sWLExsbe8fbk5KSCA8P57333qN06dJcvXqVXr164evrS2hoKACbNm3i999/v2Fo/u2334iMjGTJkiWUKVOGnJwcwsLCWL58Of379wdg+/btNG/enPr16+c7v3fv3nTv3v2G9ygmJgZXV1cAdu3axcKFC1myZAk7duxg0aJFBAcHc+DAAfbv38+aNWtuOIaIiIiIiIjIrdzT8vvPPvuMBg0aULFiRQAcHR2JjIykVq1aeHp6WgI9XHt6XaFCBb799lsAOnTowO+//87nn39e4OK3bNlChw4dKF26NABubm4sX76cRo0a3dH5a9as4Y033rA8bXdyciI0NJR169ZZjhk3bhwTJkwgIyPjnutMSUmhWLFiAIwdO5aEhAQOHTpEeHg4kZGRODs73/PYIiIiIiIi8ui6pyf1p0+fxtfX12qbu7s7ycnJ+bYD+Pr6kpKSAlx7AyAiIoLXX3+dOnXqWB134cIFgoKCLK9DQkKoVatWvu1lypRhzpw5nD59murVq1uN4enpecfzSEpKolOnTlbbPDw8uHLlCnl5eQBUq1aNDh06EBERwfjx462OXblyJQkJCZbXAwcOtLyh0LdvXzIzMzl9+jSNGzcmJCTEMn54eDivvfYaI0eOpHLlyndcr4iIiIiIiMif3VOoL1++PD///LPVtqSkJEqXLs3x48fzHZ+YmEjDhg05ceIEABUrVqRXr15MnjwZk8lkOe5ul9+XL1+ekydPWm07fPgwZrOZGjVq3HYeZcuW5fjx4zz11FOWbenp6bi4uODg8L9FDAMGDKB79+58+eWXVuffyfL7uXPnkpycTKlSpSz76tevT7FixQgICLhtjSIiIiLe3nf+0MKWY4o8aOpjMYKC9vE9hfpmzZqxZMkSunfvToUKFcjOziYiIoKGDRty5swZdu7cib//tS90+fLLL0lMTKR+/fps2fK/L3zr2bMnn332Gb/++ivdunW7p+JfeeUV3nzzTdq0aYOXlxcZGRlMnDiRN998845Cfffu3ZkwYQJ16tTB29ub7Oxspk2blq+e66sLrn/O/m4EBwfTq1cv4uLi6NGjx12fLyIiIpKaeqlQx/P29iz0MUUeNPWxGMH1Pi5IsL+nUO/h4WFZjm42m8nIyKBZs2YEBgby0ksvMX36dJYsWQKAj48PS5cuxdHR0WoMk8nE9OnTadu27W2v99fl9wAjRoygbt26jB49miFDhuDo6EhGRgadOnWiadOmdzSPmjVrMnz4cIYPH05ubi45OTm0bNnyhuG9UqVKvPbaa7z//vuWbX9dfu/n58eUKVOsznNwcGDatGn06NGDFi1aULZs2TuqTUREREREROR2TGaz2WzrIuT29C6kGIHeURcjUB+LEaiPxQjUx2IEhfGk/p6+/V5EREREREREbE+hXkRERERERMROKdSLiIiIiIiI2CmFehERERERERE7pVAvIiIiIiIiYqcU6kVERERERETs1D39Tr08WG1HbrF1CSIiIiIihS4m1N/WJYjYPT2pvw+CgoI4duyYrcsQERERERERg1OoFxEREREREbFThlx+n56ezrhx47h06RJpaWm0aNGC7du3k5CQgMlkYvLkyTRs2JCyZcsyefJk3N3dKVWqFK6urkRERNxwzP379xMZGYmTkxPFihVj9uzZAFbX6dy5M4GBgZZzLl26xLhx40hLSwNg/PjxVKtWjdDQUP7zn/+QmZlJv379aNOmzf2/KSIiIiIiImI4hgz1iYmJvPzyy7Rq1YpTp04RFBTEU089xXfffcfTTz/NN998w7hx4+jcuTMzZ87kySefJCoqilOnTt10zB07dtCyZUv69evHzp07uXjxImlpafmu8+dQv3jxYp5//nkCAwP5448/ePvtt1m2bBn79u1j48aNAOzevfu+3w8RERERkYeRt7enTc8XeRgUtI8NGepLly7N+++/zyeffIKHhwc5OTl06dKFzZs3k5qair+/P05OTpw+fZonn3wSgGeeeYaEhISbjjlw4EAWL17Ma6+9RtmyZaldu/YNr/NnR44cYe/evXz00UcAXLx4EQ8PDyZMmMCECRNIT0+nXbt29+9GiIiIiIg8xFJTL93zud7engU6X+RhcL2PCxLsDRnqY2JiqFOnDoGBgezdu5cvvviCv//978yaNYtTp04xceJEAHx8fDh69ChVqlThhx9+uOWY27Zt49VXXyUkJIQlS5awfv16Ll26lO86f1apUiXatWtH27ZtOXv2LPHx8Zw+fZpDhw7x7rvvkpmZSdOmTWnfvj1OTob8U4iIiIiIiMh9ZMgk2axZM8LCwti2bRslSpTA0dGR7OxsXnzxRfbs2cMTTzwBwKRJkxg7dixFixbF2dmZsmXL3nTMv/3tb4SGhlqOnTJlCsePH893naysLMs5AwcOZNy4caxfv5709HSGDBmCt7c3qampdOjQgaJFi9K3b18FehEREREREbknJrPZbLZ1EbayevVqWrdujZeXF1FRUTg7OzNkyBBbl3VDWlokRqBlcmIE6mMxAvWxGIH6WIxAy+8LqFSpUvTt25eiRYvi6elJREQEQ4YM4cKFC1bHeXh4EB0dbaMqRURERERERG7skX5Sb0/0LqQYgd5RFyNQH4sRqI/FCNTHYgSF8aTeoRDrEREREREREZEHSKFeRERERERExE4p1IuIiIiIiIjYKYV6ERERERERETv1SH/7vb1oO3KLrUsQERERERGxOzGh/rYu4b7Tk3oRERERERERO6Un9YUkPT2dcePGcenSJdLS0ujcuTO1atVi8uTJuLu7U6pUKVxdXYmIiCA2Npbt27djMplo06YNvXr1snX5IiIiIiIiYocU6gtJYmIiL7/8Mq1ateLUqVMEBQXh7u7OzJkzefLJJ4mKiuLUqVMcPXqUhIQE4uLiMJlM9O7dm3/84x9UqlTJ1lMQERERERERO6NQX0hKly7N+++/zyeffIKHhwc5OTmcPn2aJ598EoBnnnmGhIQEjhw5QkpKCr179wbgwoUL/Oc//1GoFxERERERKWTe3p62LuG2ClqjQn0hiYmJoU6dOgQGBrJ3716++OILfHx8OHr0KFWqVOGHH34AoFKlSlSpUoX33nsPk8nEypUrqVq1qo2rFxERERERMZ7U1Eu2LuGWvL09SU29VKBgr1BfSJo1a0ZYWBjbtm2jRIkSODo6MnHiRMaOHUvRokVxdnambNmyVK9enb///e90796drKwsateuTdmyZW1dvoiIiIiIiNghk9lsNtu6CKNavXo1rVu3xsvLi6ioKJydnRkyZMg9jfWwv8MkcieuvxMpYs/Ux2IE6mMxAvWxGIGe1D/kSpUqRd++fSlatCienp5ERETYuiQRERERERExEIX6++ill17ipZdesnUZIiIiIiIiYlAOti5ARERERERERO6NPlMvIiIiIiIiYqf0pF5ERERERETETinUi4iIiIiIiNgphXoRERERERERO6VQLyIiIiIiImKnFOpFRERERERE7JRCvYiIiIiIiIidUqi3sby8PCZOnEjXrl0JCgoiMTHRav/OnTvp2LEjXbt2Zf369Xd0jsiDdi99nJ2dzejRowkMDKRTp0589tlntihdxOJe+vi6s2fP0rRpU44dO/YgSxa5oXvt5SVLltC1a1cCAgKIj49/0GWLWLnX/20xcuRIunXrRmBgoP47WWzuTnLblStX6Natm6Vf7ynrmcWm/vWvf5lDQkLMZrPZ/H//93/mgQMHWvZlZWWZW7RoYT5//rw5MzPTHBAQYD59+vQtzxGxhXvp4w0bNpjDw8PNZrPZfO7cOXPTpk1tUbqIxb308fV9gwcPNrdq1cp89OhRm9Qu8mf30st79+41v/HGG+bc3Fxzenq6ef78+bYqX8RsNt9bH3/66afmt956y2w2m81fffWVeciQITapXeS62+W2gwcPml999VVzw4YNLf8b4l6ynp7U29j+/ftp3LgxAHXq1OGnn36y7Dt27BgVKlSgePHiuLi48Mwzz/Ddd9/d8hwRW7iXPn7ppZcYNmyY5ThHR8cHXrfIn91LHwNERkbSrVs3ypQpY5O6Rf7qXnr5q6++omrVqrz55psMHDiQF154wUbVi1xzL33s5+dHbm4ueXl5pKen4+TkZKvyRYBb9zFAVlYW7777LpUqVbrjc25EnW5j6enpeHh4WF47OjqSk5ODk5MT6enpeHp6Wva5u7uTnp5+y3NEbOFe+tjd3d1y7ltvvUVwcPCDLlvEyr308aZNm/Dy8qJx48YsXbrUFmWL5HMvvZyWlkZKSgqLFy8mOTmZQYMG8fHHH2MymWwxBZF76uOiRYty/PhxWrduTVpaGosXL7ZF6SIWt8ttzzzzzF2fcyN6Um9jHh4eZGRkWF7n5eVZ/mB/3ZeRkYGnp+ctzxGxhXvpY4ATJ078//btWDWxIIzi+NkmSRVtfAEFwVJtfYVgYSCFaOEbWIhiZSEB06eyEdTSIlWwsVAICBZWCnaWQREblahxUinZZcnCZPFyyf/XXga+gcPlHu6M0um04vG4bm5uzjs08AebHLdaLb28vCiVSmk8Hiufz2s2m519duAzmyx7vV7FYjFdXFzI7/fr8vJSi8Xi7LMDRzY5rtVqisViarfbenp6UqFQ0Nvb29lnB45sepvNGkq9wyKRiLrdriRpOBwqGAyengUCAU2nUy2XS223Ww0GA4XD4S/XAE6wyfF8Plcmk1Eul9Pt7a1TowMnNjluNptqNBqq1+sKhUKqVCry+XxObQGQZJflaDSqXq8nY4xeX1+12Wzk9Xod2gFgl+Pr6+vTjwOPx6P9fq/393dH5gekr3P8P9f8MsaY742K7zgcDiqVSppMJjLG6P7+XqPRSOv1Wnd3d+p0Onp8fJQxRolEQslk8q9rAoGA01vBD2aT43K5rOfn59/uEFWrVV1dXTm4E/xkNjn+LJVKqVQq8T6G42yz/PDwoH6/L2OMstns6U4n4ASbHK9WKxWLRc1mM+12O6XTaU4CwlH/yvHR528Im65HqQcAAAAAwKU4fg8AAAAAgEtR6gEAAAAAcClKPQAAAAAALkWpBwAAAADApSj1AAAAAAC4FKUeAAAAAACXotQDAAAAAOBSlHoAAAAAAFzqA/kGPqBdxWs2AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "best_rf = best_clf_rf.best_estimator_.fit(X_train,y_train)\n",
    "feat_importances = pd.Series(best_rf.feature_importances_, index=X_train.columns)\n",
    "plt.figure(figsize=(16,8))\n",
    "feat_importances.nlargest(30).plot(kind='barh');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "530cd499",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "50e46dba",
   "metadata": {},
   "source": [
    "### BRIEF TAKEAWAYS ON RF CLASSIFIER & DEPARTMENT CONTRIBUTIONS  [FIX REQUIRED]\n",
    "\n",
    "* Fairly expected insights:\n",
    "    * Age plays an important role in informing the email-opt in decision; this is likely related to the uptake of new technologies among different age groups. \n",
    "    * (Fruit & Veg) and (Confectionery) are heavily marketed items online and in the catalogue, and is expected to be an important \"draw card\" for customer shopping / customer behaviour influence\n",
    "    * While Tobacco was a massive contributor to overall sales earlier, as expected it doesn't play a significant role in loyalty customer behaviour; likely stemming from the fact that tobacco sales do not provide loyalty points\n",
    "    * Customer duration is important, the longer you frequent the same location, the more likely you are to subscribe to the email specials for that location - fairly logical to presume but useful to see it statistically represented here.\n",
    "    * Sales are important, customers who spend more are more likely to opt into the email marketing; conversely, the causality could in a continuous cycle, as customer's who opt into the marketing also then subsequently spend more. Either way, this isn't exactly a novel discovery \n",
    "   \n",
    "   \n",
    "* somewhat surprising takeaways: \n",
    "    * Slightly surprising to see the \"top-up\" shop items: (milk, pet food, snackfoods, softdrinks) among some of the most important features as email opt-in shoppers are expected to partake in more 'staple' purchases like (meat, deli)\n",
    "    * post code (proximity to store) is not as important as expected, the most geographically close post_code only represents about the (20th) top feature. \n",
    "    * Gender is not important in the email-marketing decision\n",
    "        * Although, this was slightly supported by the non-significant difference seen in the two groups means in the initial data exploration section"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a78dbe5d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6bd27c60",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "b875d97f",
   "metadata": {},
   "source": [
    "## TUNED XGBOOST \n",
    "\n",
    "### FIRST NARROW DOWN FEATURE SPACE WITH RANDOMIZED SEARCH\n",
    "* NOTE: even the randomized search of over the param_grid options can take a long time to run, even with a fast PC. If the code is unable to finish running, try reducing the cross_validations (CV=5) or n_iter (I lowered 1000 > 250, try even 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "f1ab8558",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 250 candidates, totalling 1250 fits\n",
      "[12:46:48] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "\n",
      " MODEL:  XGB \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.8387314439946018\n",
      "Best Parameters: {'subsample': 0.8, 'sampling_method': 'uniform', 'reg_lambda': 1.5, 'reg_alpha': 0, 'n_estimators': 1000, 'min_child_weight': 0.1, 'max_depth': 15, 'learning_rate': 0.7, 'gamma': 0, 'colsample_bytree': 1}\n"
     ]
    }
   ],
   "source": [
    "xgb = XGBClassifier(random_state=1)\n",
    "\n",
    "# different param featrues to be randomly iterated over \n",
    "param_grid = {\n",
    "    'n_estimators': [20, 50, 100, 250, 500, 1000],\n",
    "    'colsample_bytree': [0.2, 0.5, 0.7, 0.8, 1],\n",
    "    'max_depth': [2, 5, 10, 15, 20, 25, None],\n",
    "    'reg_alpha': [0, 0.5, 1],\n",
    "    'reg_lambda': [1, 1.5, 2],\n",
    "    'subsample': [0.5, 0.6, 0.7, 0.8, 0.9],\n",
    "    'learning_rate':[.01, 0.1, 0.2, 0.3, 0.5, 0.7, 0.9],\n",
    "    'gamma':[0, .01, 0.1, 1, 10, 100],\n",
    "    'min_child_weight':[0, .01, 0.1, 1, 10, 100],\n",
    "    'sampling_method': ['uniform', 'gradient_based']}\n",
    "\n",
    "# run the randomized search on param grid\n",
    "clf_xgb_rnd = RandomizedSearchCV(xgb, param_distributions = param_grid, n_iter = 250, \n",
    "                                 cv = 5, verbose = True, n_jobs = -1)\n",
    "\n",
    "# fit the best randomizedsearch_CV model and print brief summary\n",
    "best_clf_xgb_rnd = clf_xgb_rnd.fit(X_train, y_train)\n",
    "model_performance_report(best_clf_xgb_rnd,'XGB')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8f6e974d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "f320ffb5",
   "metadata": {},
   "source": [
    "## FIT A MORE REFINED XGBOOST MODEL WITH THE PREVIOUSLY OPTIMAL PARAMETERS FROM RANDOMIZED SEARCH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "a95857b2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 648 candidates, totalling 3240 fits\n",
      "[12:54:30] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "\n",
      " MODEL:  XGB \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.8441295546558705\n",
      "Best Parameters: {'colsample_bytree': 1, 'gamma': 0, 'learning_rate': 0.7, 'max_depth': 13, 'min_child_weight': 0.1, 'n_estimators': 800, 'reg_alpha': 0, 'reg_lambda': 1.2, 'sampling_method': 'uniform', 'subsample': 0.8}\n"
     ]
    }
   ],
   "source": [
    "xgb = XGBClassifier(random_state=1)\n",
    "\n",
    "# refine each parameter in a small range around the above randomized-param values\n",
    "param_grid = {\n",
    "    'n_estimators': [800, 1000, 1200],\n",
    "    'colsample_bytree': [1, 1.2],\n",
    "    'max_depth': [13, 15, 17],\n",
    "    'reg_alpha': [0],\n",
    "    'reg_lambda': [1.2, 1.5, 1.7],\n",
    "    'subsample': [0.8],\n",
    "    'learning_rate':[0.7, 0.8],\n",
    "    'gamma':[0, 0.01],\n",
    "    'min_child_weight':[0.1, 0.3, 0.5],\n",
    "    'sampling_method': ['uniform']}\n",
    "\n",
    "# fit the model from the now refined-grid params\n",
    "clf_xgb = GridSearchCV(xgb, param_grid = param_grid, cv = 5, verbose = True, n_jobs = -1)\n",
    "\n",
    "# fit best model and print brief summary\n",
    "best_clf_xgb = clf_xgb.fit(X_train, y_train)\n",
    "model_performance_report(best_clf_xgb,'XGB')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "e9ad8e8d",
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_tuned_results_dict['tuned_xgb'] = np.round(best_clf_xgb.best_score_,4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "5b45b31b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model \n",
    "pickle.dump(best_clf_xgb, open(\"xgboost_best.pickle.dat\", \"wb\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c90897d0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca23d821",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "8911efe2",
   "metadata": {},
   "source": [
    "## TUNED CLASSIFICATION MODEL RESULTS \n",
    "* Model scores on the training data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "id": "cba16886",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'tuned_logreg': 0.6355,\n",
       " 'tuned_knn': 0.7343,\n",
       " 'tuned_rfc': 0.8493,\n",
       " 'tuned_xgb': 0.8441,\n",
       " 'tuned_svc': 0.9008}"
      ]
     },
     "execution_count": 194,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_tuned_results_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "id": "a9c30557",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show values in bar graph \n",
    "plt.figure(figsize=(8, 6))\n",
    "keys = cv_tuned_results_dict.keys()\n",
    "values = cv_tuned_results_dict.values()\n",
    "plt.bar(keys, values)\n",
    "plt.xticks(rotation='45')\n",
    "plt.title(\"Tuned Classificaiton CV Model Performances (Training Data)\");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c6fcfbb",
   "metadata": {},
   "source": [
    "### BRIEF TAKEAWAYS: \n",
    "* Tuned logistic regression has a mild gain ~ 3%, not enough to be considered.\n",
    "* While tuned KNN sees a very generous increase: 58% to 73%, it still underperforms\n",
    "* RF classifier imrpoved only marginally, a little over 1% but its un-tuned performance means the model might have application in larger data sets with less computing power / resources to optimize the other models\n",
    "* XGB sees a modest gain of around 7%, but the computational burden is significantly higher than the RFC, and loses out in that respect\n",
    "* tuned SVC sees a massive gain 62% to 90%, a clear winner\n",
    "    * Although, the SVC may have problems adapting to larger data sets, more so than any other model, which affects its viability on a massive scale without further internal optimisations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b3be2d10",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "19d4ad06",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "9929d76f",
   "metadata": {},
   "source": [
    "<a id='add_vc'></a>\n",
    "\n",
    "## ADDITIONAL MODEL APPROACHES: VOTING CLASSIFIER MODIFICATIONS \n",
    "* useful tool if we are not 100% sure which classifier is right for the problem\n",
    "\n",
    "### EXPLANATION\n",
    "* Hard = popular vote, majority wins\n",
    "* Soft = predicts the winner based on the argmax of the sums of the predicted probabilities \n",
    "    * Soft is considered superior if the classification models have been prior calibrated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "id": "ab3a2627",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import a package to more easily visualise the results in table format\n",
    "from tabulate import tabulate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "id": "de4e41dc",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define our best estimators from each model for clarity \n",
    "best_lr = best_clf_lr.best_estimator_     # best logistic regression \n",
    "best_knn = best_clf_knn.best_estimator_   # best k-nearest neighbors\n",
    "best_svc = best_clf_svc.best_estimator_   # best support vector classifier  \n",
    "best_rf = best_clf_rf.best_estimator_     # best random forest \n",
    "best_xgb = best_clf_xgb.best_estimator_   # best XGBoost"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "347c2dca",
   "metadata": {},
   "source": [
    "* define the different iterations of voting classifiers\n",
    "    * soft vs hard\n",
    "    * sub-sample of models vs all models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "id": "7d061201",
   "metadata": {},
   "outputs": [],
   "source": [
    "voting_clf_soft = VotingClassifier(estimators = [('knn',best_knn),('rf',best_rf),('svc',best_svc)], \n",
    "                                   voting = 'soft') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "id": "f7bd2158",
   "metadata": {},
   "outputs": [],
   "source": [
    "voting_clf_hard = VotingClassifier(estimators = [('knn',best_knn),('rf',best_rf),('svc',best_svc)], \n",
    "                                   voting = 'hard') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "id": "2c0c2e38",
   "metadata": {},
   "outputs": [],
   "source": [
    "voting_clf_all = VotingClassifier(estimators = [('knn',best_knn),('rf',best_rf),('svc',best_svc), ('lr', best_lr)], \n",
    "                                  voting = 'soft') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "id": "a2c7e040",
   "metadata": {},
   "outputs": [],
   "source": [
    "voting_clf_xgb = VotingClassifier(estimators = [('knn',best_knn),('rf',best_rf),('svc',best_svc), ('xgb', best_xgb),\n",
    "                                                ('lr', best_lr)], \n",
    "                                  voting = 'soft')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "id": "faa16810",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[13:08:32] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[13:08:34] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[13:08:35] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[13:08:36] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "[13:08:38] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "+-------------------+--------------------+\n",
      "| VOTING CLASSIFIER |  MODEL 5-CV MEAN   |\n",
      "+-------------------+--------------------+\n",
      "| VOTING CLF SOFT:  | 0.843859649122807  |\n",
      "| VOTING CLF HARD:  | 0.875033738191633  |\n",
      "|  VOTING CLF ALL:  | 0.8489878542510121 |\n",
      "|  VOTING CLF XBG:  | 0.8697705802968961 |\n",
      "+-------------------+--------------------+\n"
     ]
    }
   ],
   "source": [
    "print(tabulate([['VOTING CLF SOFT:', cross_val_score(voting_clf_soft,X_train,y_train,cv=5).mean()],\n",
    "              ['VOTING CLF HARD:', cross_val_score(voting_clf_hard,X_train,y_train,cv=5).mean()],\n",
    "                ['VOTING CLF ALL: ', cross_val_score(voting_clf_all,X_train,y_train,cv=5).mean()], \n",
    "                ['VOTING CLF XBG: ',cross_val_score(voting_clf_xgb,X_train,y_train,cv=5).mean()]],\n",
    "               headers=['VOTING CLASSIFIER','MODEL 5-CV MEAN'],\n",
    "                tablefmt='pretty', numalign='center'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a49b75ab",
   "metadata": {},
   "source": [
    "* Conversely, the hard classifier outperforms the soft classifier with the same 3 models\n",
    "    * this might be addressed by influencing the weights of each model within the voter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a6d4430b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "0b537e46",
   "metadata": {},
   "source": [
    "### EXPERIMENT WIHT USING DIFFERENT MODEL WEIGHTS IN THE SOFT VOTING CLASSIFIER \n",
    "* A grid search will iterate through different weightings \n",
    "\n",
    "#### TRY EXPERIMENT WITH DIFFERENT WEIGHTS ON THE TWO BEST SOFT VOTING CLASSIFIERS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "0175caf5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialise dictionary of param weights \n",
    "params_soft_vc = {'weights' : [[1,1,1],[1,1,2],[1,2,1],[1,2,2],\n",
    "                       [2,1,1],[2,2,1],[2,1,2]]}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "id": "929c1eea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 7 candidates, totalling 35 fits\n",
      "\n",
      " MODEL:  VC-soft Weights \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.8748987854251012\n",
      "Best Parameters: {'weights': [1, 1, 2]}\n"
     ]
    }
   ],
   "source": [
    "# pass in the weights dict as the param grid \n",
    "vote_weights = GridSearchCV(voting_clf_soft, param_grid=params_soft_vc, \n",
    "                           cv = 5, verbose = True, n_jobs = -1)\n",
    "\n",
    "# fit the best model\n",
    "best_clf_weight = vote_weights.fit(X_train, y_train)\n",
    "\n",
    "# print results\n",
    "model_performance_report(best_clf_weight,'VC-soft Weights')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a38a2a2",
   "metadata": {},
   "source": [
    "* adjusting the weights to further weight the SVC results in a higher score (unsurprising, as the SVC alone has a higher score than the voting classifier) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "63a0d752",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "36cdd8d7",
   "metadata": {},
   "source": [
    "## Experiment with different weightings for the 5 models\n",
    "* Try iterations with more weight on the models we know performed best on the single model cross-validation steps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "986bd0c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "params_soft_xgb = {'weights' : [[1,1,1,1,1],[1,1,2,1,1],[1,1,2,2,1],[1,2,2,2,1]]}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "id": "d788d6f5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 4 candidates, totalling 20 fits\n",
      "[13:17:14] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "\n",
      " MODEL:  VC-soft Weights: XGB \n",
      " ------------------------------ \n",
      "\n",
      "Best Score: 0.8855600539811066\n",
      "Best Parameters: {'weights': [1, 1, 2, 2, 1]}\n"
     ]
    }
   ],
   "source": [
    "# pass in the weights dict as the param grid \n",
    "vote_weights = GridSearchCV(voting_clf_xgb, param_grid=params_soft_xgb, \n",
    "                           cv = 5, verbose = True, n_jobs = -1)\n",
    "\n",
    "# fit the best model\n",
    "best_clf_weight = vote_weights.fit(X_train, y_train)\n",
    "\n",
    "# print results\n",
    "model_performance_report(best_clf_weight,'VC-soft Weights: XGB')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d1aa97e2",
   "metadata": {},
   "source": [
    "* Not entirely surprising that the model with extra weight on the: SVC and XGBoost performed better, given these models outperformed individually as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5d8c5f0c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "17226d0d",
   "metadata": {},
   "source": [
    "### BRIEF TAKEAWAYS ON THE VOTING CLASSIFIER\n",
    "* Somewhat unsurprisingly, the models with extra weight on the already outperforming models perform better. \n",
    "* Given the additional complexities of forecasting the test data set with a voting classifier and the fact it is not essentially providing any additional \"value\" by over-weighting the model we already know performs best, we will perform test-data predictions with the individual tuned models only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "02318ad2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "2364c480",
   "metadata": {},
   "source": [
    "<a id='oos_models'></a>\n",
    "\n",
    "## OUT OF SAMPLE (TEST-DF) PREDICTIONS FROM TUNED MODELS\n",
    "* Use the tuned model from the training data to predict outcomes from the test data for each of the tuned models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "id": "4eb64e29",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialise a dictionary for these test data results\n",
    "tuned_model_test_results = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "id": "6ea69447",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define our best estimators from each model: \n",
    "best_lr = best_clf_lr.best_estimator_     # best logistic regression \n",
    "best_knn = best_clf_knn.best_estimator_   # best k-nearest neighbors\n",
    "best_svc = best_clf_svc.best_estimator_   # best support vector classifier  \n",
    "best_rf = best_clf_rf.best_estimator_     # best random forest \n",
    "best_xgb = best_clf_xgb.best_estimator_   # best XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "1efd7a77",
   "metadata": {},
   "outputs": [],
   "source": [
    "# test model on data inputs of x_test to get results of y_test\n",
    "log_reg_pred = best_lr.predict(X_test).astype(int)\n",
    "knn_pred = best_knn.predict(X_test).astype(int)\n",
    "svc_pred = best_svc.predict(X_test).astype(int) \n",
    "rf_pred = best_rf.predict(X_test).astype(int) \n",
    "xbg_pred = best_xgb.predict(X_test).astype(int) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "id": "a8096cdf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# append the test-df model accuracy results (using: roc_auc score)\n",
    "tuned_model_test_results['tuned_lr'] = np.round(roc_auc_score(y_test, log_reg_pred),4)\n",
    "tuned_model_test_results['tuned_knn'] = np.round(roc_auc_score(y_test, knn_pred),4)\n",
    "tuned_model_test_results['tuned_svc'] = np.round(roc_auc_score(y_test, svc_pred),4)\n",
    "tuned_model_test_results['tuned_rf'] = np.round(roc_auc_score(y_test, rf_pred),4)\n",
    "tuned_model_test_results['tuned_xgb'] = np.round(roc_auc_score(y_test, xbg_pred),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a4ceaf7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "d203d2fc",
   "metadata": {},
   "source": [
    "<a id='model_results'></a>\n",
    "\n",
    "## MODEL RESULTS SUMMARY AND DISCUSSION\n",
    "\n",
    "* Interestingly enough, (Ex. logistic regression) each of the models performed better on the test data than on the training data. \n",
    "* The RFC gained almost 10% accuracy and the SVC around 5%, this might suggests that the models were not overly fitted to the training data and actually have predictive capacity in correctly classifying new data\n",
    "    * One avenue to explore is whether the upscaling of the minority class has contributed to this in any way, thus, the total data set should be recalibrated in the future and different techniques of class balancing should be tried to ensure the results continue to hold true."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "id": "e7f40b6e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'tuned_lr': 0.625,\n",
       " 'tuned_knn': 0.75,\n",
       " 'tuned_svc': 0.9583,\n",
       " 'tuned_rf': 0.9583,\n",
       " 'tuned_xgb': 0.875}"
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# show results \n",
    "tuned_model_test_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2456fb0e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "id": "b6c165c5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show values in bar graph \n",
    "plt.figure(figsize=(9, 7))\n",
    "keys = tuned_model_test_results.keys()\n",
    "values = tuned_model_test_results.values()\n",
    "plt.bar(keys, values)\n",
    "plt.xticks(rotation='45')\n",
    "plt.title(\"Tuned Classificaiton CV Model Performances (Test Data: ROC_AUC SCORES)\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e4c7f64d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3ab2a973",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3c7bbd35",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "36b0f8a6",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}