clean_code_models.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.cross_validation import train_test_split, KFold\n",
    "from sklearn.metrics import f1_score, precision_score, recall_score, accuracy_score, confusion_matrix\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from itertools import izip\n",
    "from math import exp, log\n",
    "from unbalanced_dataset.over_sampling import SMOTE\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.svm import SVC\n",
    "from statsmodels.discrete.discrete_model import Logit\n",
    "from statsmodels.tools import add_constant\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "from sklearn.preprocessing import scale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 545,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def preprocess(dropcols=['last_sign_in_at'] ,today='2016-06-20'):\n",
    "    # Defining today's day for reference\n",
    "    date = pd.to_datetime(today)\n",
    "    \n",
    "    # Loading csv extracted from database into a Pandas DataFrame\n",
    "    df = pd.read_csv('dataset.csv', usecols=range(1,17))\n",
    "    \n",
    "    # Dropping Vango's team ids\n",
    "    vango_ids = [38175, 1, 1326, 587, 736, 45651, 67966, 48516, 84261, 30975, 4260]\n",
    "    for _id in vango_ids:\n",
    "        df = df[df.id != _id]\n",
    "    \n",
    "    # Treating date fields and converting them to datetime timestamp\n",
    "    df.created_at = pd.to_datetime(df['created_at'])\n",
    "    df.last_session = pd.to_datetime(df['last_session'])\n",
    "    df.last_favorited_artwork_date = pd.to_datetime(df['last_favorited_artwork_date'])\n",
    "    df.last_followed_artist_date = pd.to_datetime(df['last_followed_artist_date'])\n",
    "    \n",
    "    # Creating function to extract only the days (int) from TimeDelta objects\n",
    "    def extract_days(x):\n",
    "        try:\n",
    "            return x.days\n",
    "        except:\n",
    "            return 0\n",
    "    \n",
    "    # New feature: Extracting the difference between last session and today\n",
    "    df['days_from_last_session'] = date - df.last_session\n",
    "    df['days_from_last_session'] = df.days_from_last_session.apply(lambda x: extract_days(x))\n",
    "\n",
    "    # New feature: Extracting the difference between last session and when user registered\n",
    "    df['diff_created_to_last'] = df.last_session - df.created_at\n",
    "    df['diff_created_to_last'] = df.diff_created_to_last.apply(lambda x: extract_days(x))\n",
    "\n",
    "    # New feature: Extracting the difference between last favorited artwork and today\n",
    "    df['diff_last_fav_artwork_to_today'] = date - df.last_favorited_artwork_date\n",
    "    df['diff_last_fav_artwork_to_today'] = df.diff_last_fav_artwork_to_today.apply(lambda x: extract_days(x))\n",
    "\n",
    "    # New feature: Extracting the difference between last artists followed and today\n",
    "    df['diff_last_artist_folw_to_today'] = date - df.last_followed_artist_date\n",
    "    df['diff_last_artist_folw_to_today'] = df.diff_last_artist_folw_to_today.apply(lambda x: extract_days(x))\n",
    "    \n",
    "    df = df.set_index('id')\n",
    "\n",
    "    # Dropping columns\n",
    "    df = df.drop(drop_cols, axis=1)\n",
    "    \n",
    "    # Filling NaN values with unknown for categorical variables before dummifying\n",
    "    df.gender.fillna('unknown', inplace=True)\n",
    "    df.user_type.fillna('unknown', inplace=True)\n",
    "    df.os.fillna('unknown', inplace=True)\n",
    "\n",
    "    # Dropping about 100 recent user ids missing all information\n",
    "    df.num_sessions.dropna(inplace=True)\n",
    "\n",
    "    # Dropping 3 records based on gender feature\n",
    "    df = df[(df.gender != 'male (hidden)') & (df.gender != 'female (hidden)')]\n",
    "    \n",
    "    # Dummifying categorical variables (gender, user_type, and os)\n",
    "    df = pd.concat([df, pd.get_dummies(df.gender, prefix='gender', drop_first=True)], axis=1)\n",
    "    df = df.drop('gender', axis=1)\n",
    "\n",
    "#     df = pd.concat([df, pd.get_dummies(df.user_type, prefix='user_type', drop_first=True)], axis=1)\n",
    "    df = df.drop('user_type', axis=1)\n",
    "\n",
    "#     df = pd.concat([df, pd.get_dummies(df.os, prefix='os', drop_first=True)], axis=1)\n",
    "    df = df.drop('os', axis=1)\n",
    "\n",
    "    # Finally dropping off random Na values (101 records)\n",
    "    df.dropna(inplace=True)\n",
    "    \n",
    "    return df\n",
    "\n",
    "def create_label(df, name='label', original_col_name='diff_created_to_last', smaller_than=200):\n",
    "    df[name] = np.where(df[original_col_name] < smaller_than, 1, 0)\n",
    "    \n",
    "    # Dropping date columns due to high relationship and used to create the label:\n",
    "    df.drop('diff_last_fav_artwork_to_today', axis=1, inplace=True)\n",
    "    df.drop('diff_last_artist_folw_to_today', axis=1, inplace=True)\n",
    "    df.drop('diff_created_to_last', axis=1, inplace=True)\n",
    "    df.drop('days_from_last_session', axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 546,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def standard_confusion_matrix(y_true, y_predict):\n",
    "    [[tn, fp], [fn, tp]] = confusion_matrix(y_true, y_predict)\n",
    "    return np.array([[tp, fp], [fn, tn]])\n",
    "    \n",
    "    \n",
    "def roc_curve(probabilities, labels):\n",
    "    thresholds = np.sort(probabilities)\n",
    "\n",
    "    tprs = []\n",
    "    fprs = []\n",
    "\n",
    "    num_positive_cases = sum(labels)\n",
    "    num_negative_cases = len(labels) - num_positive_cases\n",
    "\n",
    "    for threshold in thresholds:\n",
    "        # With this threshold, give the prediction of each instance\n",
    "        predicted_positive = probabilities >= threshold\n",
    "        # Calculate the number of correctly predicted positive cases\n",
    "        true_positives = np.sum(predicted_positive * labels)\n",
    "        # Calculate the number of incorrectly predicted positive cases\n",
    "        false_positives = np.sum(predicted_positive) - true_positives\n",
    "        # Calculate the True Positive Rate\n",
    "        tpr = true_positives / float(num_positive_cases)\n",
    "        # Calculate the False Positive Rate\n",
    "        fpr = false_positives / float(num_negative_cases)\n",
    "\n",
    "        fprs.append(fpr)\n",
    "        tprs.append(tpr)\n",
    "\n",
    "    return tprs, fprs, thresholds.tolist()\n",
    "\n",
    "\n",
    "def fit_logistic_regression(X, y):\n",
    "    \n",
    "    sm = SMOTE(kind='regular')\n",
    "    X_resampled, y_resampled = sm.fit_transform(X, y)\n",
    "    \n",
    "    # Splitting train and test data\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3)\n",
    "    \n",
    "    # Fitting regression and getting its scores\n",
    "    log_reg = LogisticRegression()\n",
    "    log_reg.fit(X_train, y_train)\n",
    "    predict_log = log_reg.predict(X_test)\n",
    "    print \"\\nLogistic Regression Scores:\\n\"\n",
    "    print \"Accuracy on test set:\", log_reg.score(X_test, y_test)\n",
    "    print \"Precision:\", precision_score(y_test, predict_log)\n",
    "    print \"Recall:\", recall_score(y_test, predict_log)\n",
    "    \n",
    "    # Fitting multiple k-fold cross validations and getting mean scores\n",
    "    kfold = KFold(len(y))\n",
    "\n",
    "    accuracies = []\n",
    "    precisions = []\n",
    "    recalls = []\n",
    "\n",
    "    for train_index, test_index in kfold:\n",
    "        model = LogisticRegression()\n",
    "        model.fit(X[train_index], y[train_index])\n",
    "        y_predict = model.predict(X[test_index])\n",
    "        y_true = y[test_index]\n",
    "        accuracies.append(accuracy_score(y_true, y_predict))\n",
    "        precisions.append(precision_score(y_true, y_predict))\n",
    "        recalls.append(recall_score(y_true, y_predict))\n",
    "\n",
    "    print \"\\nK-Fold Cross Validation on Logistic Regression Scores:\\n\"\n",
    "    print \"accuracy:\", np.average(accuracies)\n",
    "    print \"precision:\", np.average(precisions)\n",
    "    print \"recall:\", np.average(recalls)\n",
    "    \n",
    "    cols = list(df.columns)\n",
    "    \n",
    "    print\n",
    "    print \"Beta scores:\"\n",
    "    for name, coef in izip(df.columns, model.coef_[0]):\n",
    "        print \"%s: %.4f\" % (name, coef)\n",
    "    \n",
    "    y_predict = log_reg.predict(X_test)\n",
    "    y_proba = log_reg.predict_proba(X_test)\n",
    "    cm = standard_confusion_matrix(y_test, y_predict)\n",
    "\n",
    "    tpr, fpr, thres = roc_curve(y_proba[:,0:1].flatten(), y_test)\n",
    "    plt.plot(tpr, fpr)\n",
    "    plt.show()\n",
    "    \n",
    "    fix, ax = plt.subplots(figsize=(10, 7))\n",
    "    sns.heatmap(cm, annot=True,  fmt='', square=True, \\\n",
    "                            xticklabels=['1', '0'], \\\n",
    "                            yticklabels=['1', '0']);\n",
    "    \n",
    "    print\n",
    "    print \"Likelihoods:\"\n",
    "    for i, coef in enumerate(log_reg.coef_[0]):\n",
    "#         print \"beta %s: %.5f\" % (cols[i], exp(coef))\n",
    "        if coef <0:\n",
    "            print \"*Increasing the %s by 1 point decreases the chance of label=1 by a factor of %.4f.*\" % (cols[i], exp(coef))\n",
    "        else:\n",
    "            print \"*Increasing the %s by 1 point increases the chance of label=1 by a factor of %.4f.*\" % (cols[i], exp(coef))\n",
    "        print\n",
    "        \n",
    "    print \"To double:\"\n",
    "    for i, coef in enumerate(model.coef_[0]):\n",
    "#     print \"beta %s: %.5f\" % (cols[i], log(2) / coef)\n",
    "        if coef < 0:\n",
    "            print \"*Decreasing the %s score by %d points doubles the chance of label=1.*\" % (cols[i], log(2) / coef)\n",
    "        else:\n",
    "            print \"*Increasing the %s score by %d points doubles the chance of label=1.*\" % (cols[i], log(2) / coef)\n",
    "        print"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 1: Logistic Regression with Label = activity in last 200 days"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "\n",
    "create_label(df, name='engagement200', original_col_name='diff_created_to_last', smaller_than=200)\n",
    "\n",
    "# Defining y label and X matrix\n",
    "y = df.pop('engagement200').values\n",
    "X = df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({1: 48224, 0: 9593})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 38631 new samples ...\n",
      "done!\n",
      "\n",
      "Logistic Regression Scores:\n",
      "\n",
      "Accuracy on test set: 0.676308968377\n",
      "Precision: 0.641019751952\n",
      "Recall: 0.781902227203\n",
      "\n",
      "K-Fold Cross Validation on Logistic Regression Scores:\n",
      "\n",
      "accuracy: 0.837955342251\n",
      "precision: 0.846670506015\n",
      "recall: 0.984896691247\n",
      "\n",
      "Beta scores:\n",
      "num_sessions: -0.0232\n",
      "total_artists_followed: -0.0455\n",
      "total_artworks_favorited: -0.0119\n",
      "total_artworks_shared: 0.0454\n",
      "gender_male: -0.0694\n",
      "gender_unknown: -0.7078\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAegAAAFVCAYAAAAkBHynAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwpAKWkREJAz9fwBYFQi5DN6dAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115a47710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Likelihoods:\n",
      "*Increasing the num_sessions by 1 point decreases the chance of label=1 by a factor of 0.9704.*\n",
      "\n",
      "*Increasing the total_artists_followed by 1 point decreases the chance of label=1 by a factor of 0.9698.*\n",
      "\n",
      "*Increasing the total_artworks_favorited by 1 point decreases the chance of label=1 by a factor of 0.9822.*\n",
      "\n",
      "*Increasing the total_artworks_shared by 1 point increases the chance of label=1 by a factor of 1.0851.*\n",
      "\n",
      "*Increasing the gender_male by 1 point decreases the chance of label=1 by a factor of 0.9413.*\n",
      "\n",
      "*Increasing the gender_unknown by 1 point decreases the chance of label=1 by a factor of 0.4914.*\n",
      "\n",
      "To double:\n",
      "*Decreasing the num_sessions score by -29 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the total_artists_followed score by -15 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the total_artworks_favorited score by -58 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the total_artworks_shared score by 15 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the gender_male score by -9 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the gender_unknown score by 0 points doubles the chance of label=1.*\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfcAAAGkCAYAAAAsb2x+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x114db2a10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fit_logistic_regression(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 2: Logistic Regression with Label = activity in last 15 days"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "\n",
    "create_label(df, name='engagement15', original_col_name='diff_created_to_last', smaller_than=15)\n",
    "\n",
    "# Defining y label and X matrix\n",
    "y = df.pop('engagement15').values\n",
    "X = df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({0: 31718, 1: 26109})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 5609 new samples ...\n",
      "done!\n",
      "\n",
      "Logistic Regression Scores:\n",
      "\n",
      "Accuracy on test set: 0.739898061058\n",
      "Precision: 0.698328332619\n",
      "Recall: 0.850579513418\n",
      "\n",
      "K-Fold Cross Validation on Logistic Regression Scores:\n",
      "\n",
      "accuracy: 0.735193020637\n",
      "precision: 0.674597078614\n",
      "recall: 0.803656376634\n",
      "\n",
      "Beta scores:\n",
      "num_sessions: -0.1998\n",
      "total_artists_followed: -0.0663\n",
      "total_artworks_favorited: -0.0147\n",
      "total_artworks_shared: 0.0989\n",
      "gender_male: -0.0694\n",
      "gender_unknown: -0.7463\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAegAAAFVCAYAAAAkBHynAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x114db21d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Likelihoods:\n",
      "*Increasing the num_sessions by 1 point decreases the chance of label=1 by a factor of 0.8240.*\n",
      "\n",
      "*Increasing the total_artists_followed by 1 point decreases the chance of label=1 by a factor of 0.9649.*\n",
      "\n",
      "*Increasing the total_artworks_favorited by 1 point decreases the chance of label=1 by a factor of 0.9858.*\n",
      "\n",
      "*Increasing the total_artworks_shared by 1 point increases the chance of label=1 by a factor of 1.0767.*\n",
      "\n",
      "*Increasing the gender_male by 1 point decreases the chance of label=1 by a factor of 0.8928.*\n",
      "\n",
      "*Increasing the gender_unknown by 1 point decreases the chance of label=1 by a factor of 0.5183.*\n",
      "\n",
      "To double:\n",
      "*Decreasing the num_sessions score by -3 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the total_artists_followed score by -10 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the total_artworks_favorited score by -47 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the total_artworks_shared score by 7 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the gender_male score by -9 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the gender_unknown score by 0 points doubles the chance of label=1.*\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfEAAAGkCAYAAAAhcRw5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAACb+n/vter9IozbjwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x114bc0a50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fit_logistic_regression(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 3: Logistic Regression with Label = activity in last 30 days (ignoring 0 days of activity -> 17k records dropped)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "\n",
    "def negative_to_zero(x):\n",
    "    if x == -1:\n",
    "        return 0\n",
    "    else:\n",
    "        return x\n",
    "\n",
    "df.diff_created_to_last = df.diff_created_to_last.apply(lambda x: negative_to_zero(x))\n",
    "\n",
    "df = df[df.diff_created_to_last > 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "create_label(df, name='engagement30', original_col_name='diff_created_to_last', smaller_than=30)\n",
    "\n",
    "# Defining y label and X matrix\n",
    "y = df.pop('engagement30').values\n",
    "X = df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({0: 27443, 1: 13224})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 14219 new samples ...\n",
      "done!\n",
      "\n",
      "Logistic Regression Scores:\n",
      "\n",
      "Accuracy on test set: 0.652435321268\n",
      "Precision: 0.611503531786\n",
      "Recall: 0.817512877116\n",
      "\n",
      "K-Fold Cross Validation on Logistic Regression Scores:\n",
      "\n",
      "accuracy: 0.681066998172\n",
      "precision: 0.574449831018\n",
      "recall: 0.129059241002\n",
      "\n",
      "Beta scores:\n",
      "num_sessions: -0.0710\n",
      "total_artists_followed: -0.0315\n",
      "total_artworks_favorited: -0.0099\n",
      "total_artworks_shared: 0.0649\n",
      "gender_male: -0.1417\n",
      "gender_unknown: -0.4610\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAegAAAFVCAYAAAAkBHynAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl0lfWB//H3zXKz7xsQsocQ1gCxuGCUKqnUHQVFnTAd\nZ5xOpzPTqfZ3ZjpzRDpzHFr7s+dMB53O1Dn601bpaN2auiACKpvSQICAJCSEEAIJ2W+Wm9zc3Of3\nR/C2FM2V5d7nLp/XOT3l3uea5+NXkk+e7fu1GIZhICIiIn4lzOwAIiIicj4VtIiIiB9SQYuIiPgh\nFbSIiIgfUkGLiIj4IRW0iIiIH/pSBb1//36qqqrOe3/Lli2sXLmS1atX8/LLL1/2cCIiIqEqwtMH\nnnnmGd544w3i4uLOed/pdPLDH/6QV199laioKO677z5uvPFGUlNTvRZWREQkVHg8gs7Ly+Opp546\n7/2mpiby8vKIj48nMjKS8vJy9uzZ45WQIiIiocZjQVdWVhIeHn7e+4ODgyQkJLhfx8XFMTAwcHnT\niYiIhCiPp7i/SHx8PIODg+7XQ0NDJCYmevznDMPAYrFc7G5FRET8zujYOL/7tIO6pi62156ib3DU\nvS1vSgIb/s8NF/w1v3RB//GU3UVFRbS0tGCz2YiOjmbPnj38+Z//ucevY7FY6OzUkbY3ZWQkaIx9\nQOPsfRpj79MYXzjDMOjqH6GprZ+mUzaa2vppPTPIuGuiJyPCLSyZN4WyonRKcpNJjLVe1H6+dEF/\ndtRbXV2N3W5n1apVfP/73+fBBx/EMAxWrVpFZmbmRYUQERHxV6Nj4xw/bXOXcdMpG7Yhh3t7RLiF\n/CkJFE5Loig7kYUz0omMOP/S8IWymLGalX5b8y79RuwbGmfv0xh7n8b48/XYRtjf1M3+xi4+bell\nzOlyb0tJiKIoO4niaYkUZieRlxXvsZAzMhIm3f55LvoatIiISLAwDIMzfXYON/ew+3AHR0/2u7dl\nZ8QxJz+V4uwkCqclkpoY7ZNMKmgREQk5jrFxjrcP0NTWT+PZ/w0Mj7m3l+Qkc8XMDBYUp5OeHGNK\nRhW0iIgEvb7BURpP/r6MW9oH3Dd1AaQlRrF4ViZF2UnMyktheka8iWknqKBFRCTouFwGDa191DR0\nsr+xi67+Efe28DALuVkJFGcnUTw9iSIfnra+ECpoEREJCl19dlo6Bjl4rIt9R7vcp6xjoyJYUJxO\nUXYixdlJ5E9NJCry0u+y9jYVtIiIBKzao10cbeujtWOQuuYe9/uJcVaWLsymvCSDmbnJRIQH3uKN\nKmgREQkojrFxGlr7eHPncRr/4G7r6RlxXDk7ixnTkynOTiIsLLBnrVRBi4iIX3O5DE52DlLf2jdx\nXbm+070tLTGKW67OZ2FJBomxkUE1lbQKWkRE/Ipz3MXx9gEazhby0ZN92EfH3dtTEqKYMT2JK2dn\nUVaUHvBHyl9EBS0iIqYxDIPOszd3tbQPcOxUP8dO2XD8wcxdWSkxlM9MZmZOMiU5yaQnRQfVkfIX\nUUGLiIhPuAyDM712jrfbONE+yPF2Gy0dg9hHne7PWIDsjHhKcpIoOVvIyfFR5oU2kQpaRES8Ymhk\njAON3RxvH6ClY4ATHQOMOH5/qtoCZKXGMr8ojbysBPKmJJCXFU9sdKR5of2IClpERC6bQfsYR0/2\n8fHhDvYd7XIvMmGxwNS0OHcR509JICcznpgo1dAX0ciIiMhFGx4Z40BTN7/a0ggW6B/8/TKMmckx\nVJRNZWZOCjmZ8URZ/X9yEH+ighYRkQt2qmuIbbVt7DjYfs415NzMeOYUpLJgRjpF2UmEhcDNXN6i\nghYREY/6B0fZ39TNB7WnsA056LZNzG0dZQ3npsU5XFGaSW6m53WR5ctTQYuIyHkMw6D1zCD7G7uo\nbeym+bTtnO0zpiexdGE2XynNDMhpNAOBClpERACwDTs42trH4ZZe9jd20WMbBSDMYqE0N5kFxemU\nFaeTmRITEs8hm00FLSISonoHRvn0ZD97DrXT0NrHqa4h97bYqAiunJ3FguJ05hWm6tEnE6igRURC\nRFe/nfoTE9Nn1rf2cabX7t5mjQxjTn4KJTnJzMxNoXBaok5dm0wFLSIShIyzs3bVt/adLeVeus+e\nsgaIiQpnflEai0qzyE6LIS8rQYXsZ1TQIiJBwD7qpL1nmOOnbROl3Np3zjPJcdERLJyRzszcFGbm\nJJOTGU9YmIWMjAQ6OwdMTC5fRAUtIhKADMOgq3+E5tM2ahu72H2o45ztiXFWvlKayczcifmsp6XH\n6ZnkAKOCFhEJID22EbYfOM3WfW30D517hDw7P5XZ+SnMzE0hS3daBzwVtIiIn+uxjbBpTyt7jpyh\nd2DiOnKUNZzymRkUTE1kdn4KuVkJOkIOMipoERE/NO5y8fLWJjbtaXW/Fx5mYV5hGvOL0rhm7hQt\nNBHk9F9XRMSPjDnH2XWog7d3t9Bx9jGo7PQ4biifzjVzpxAVqak0Q4UKWkTEDwzax9i69yTv723D\nNuQgPGxi9q67ri+iODvJ7HhiAhW0iIgJnOMuWjoGaDgx8UjUkZZeHE4XMVHhfP3KXG4sn05qYrTZ\nMcVEKmgRER9wjrs43j5A/Yle6k/0cbStn1HHuHv71LRYri+bRkXZNF1bFkAFLSLiFWNOF82nbROF\n3NpHY1s/jjGXe/vUtFj3pCElOcmkJESZmFb8kQpaROQyGHOOc+yUjfqzp6wb2/oZc/6+kLPT4yjJ\nTaY0d2K+66Q4q4lpJRCooEVELlJ3/wjbD57mSEsvTadsOMd/X8jTM+KZmZs8cYScm0xirApZLowK\nWkTkAnX12fnt7ha2HzjNuMvAAuRkxk+csj47tWZ8jJZnlEujghYR+ZLO9A5TvauFXXXtjLsMslJi\nuOXqfBaWpBOn9ZLlMlNBi4h40N4zzG93HmfXoQ5chsHUtFhuuyafxbOyCAvT9JriHSpoEZE/4nIZ\nNLb1c7p7iI8OnKb5tA3DmLjR67Yl+VwxM1PFLF6nghYR+QNdfXY2bmlkb0On+734mEjW3DSTRTMz\ntCCF+IwKWkQEGHWM86utjWzb1wZAyfQkrpydRUluCtPSYrV0o/icClpEQpZhGOxt6OKjA6eob+1j\n1DFOmMXCHRUFLF+cQ2SEFqYQ86igRSTkuFwGtY1dvLW7hWOnbABkpcZy5axMbiyfToKeWRY/oIIW\nkZDR1W9nb0MX79e00tk3AkD+lARuuTqf8pkZJqcTOZcKWkSCXvNpGz97o85dygBL5k1h+eJcsjPi\nTUwm8sVU0CIStPqHHDy5sZaTnYPAxAIV84vSWDwri4KpiSanE5mcClpEgk7vwCjbD5xi055Whkac\nANx7QzE3LMrWjV8SMFTQIhIUxl0uDjR28+H+Uxw41o1hQExUOCuuK+SWq/P0/LIEHBW0iAS07v4R\n6pq72VxzkrbOIWDixq/ryqZx5ewsYqL0Y04Ck/7mikhAGXWMc+REL4eae6hr7qG9Z9i9bUFxOndW\nFJCblWBiQpHLQwUtIn7NZRi0dgxy6HgPdce6OXqyn3GXAUBUZDhlRWnMLUxjTkEqU1JjTU4rcvmo\noEXE7/QPjnLoeA+NpxrYe6QD2/CYe1teVgJzC1OZk59K8fQkIsLDTEwq4j0qaBHxCy3tA2yuaeXg\nsR5sQw73+0nxVpbMncKcglRmF6SSqFm+JESooEXENM5xF3sbOtlcc5LGk/0AJMdbmV+URmluChXl\nOcSGo4UqJCSpoEXE5zp6htlRd5rtB07TNzhxtDy3MJVl5dOZW5jmfiQqIyOBzs4BM6OKmEYFLSI+\nMTDs4EBTN4eP97DrUAcA1sgwlpVP54by6brBS+SPeCxowzBYt24d9fX1WK1WHn/8cXJyctzb33zz\nTZ577jnCw8O56667uO+++7waWEQCx6hjnH1HO9l9uINDzT3uu6+T4qwsu2I6NyyarueURb6Ax++M\nzZs343A42LhxI/v372f9+vU8/fTT7u1PPPEEb7/9NtHR0dxyyy3ceuutJCToGUSRUNbSPsBvdx3n\n4LEeRsfGgYm7rxfPymROQSrTM+M1s5eIBx4LuqamhoqKCgDKysqoq6s7Z3tpaSn9/f3umzh0M4dI\naOrqs7PvaBe7D3fQfNrmfv/Wa/K5anYW09LjTEwnEng8FvTg4OA5R8QRERG4XC7CwiaePZwxYwZ3\n3303sbGxVFZWEh/veem2jAwdYXubxtg3NM7QNzDKS5uOsOnjEzjHXVgs8JXZWdx8TQELZ2YSHnZp\nv7RrjL1PY+yfPBZ0fHw8Q0ND7td/WM719fVs27aNLVu2EBsby/e+9z3effddbrrppkm/pu7K9C7d\n+eoboTzOPbYRdta1c6i5h/rWPgCireHce0MJZcVppCfFTHyue/CS9hPKY+wrGmPfuJhfgjwW9KJF\ni9i6dSvLly+ntraWkpIS97aEhARiYmKwWq1YLBZSU1Ox2WyTfDURCWQ19Wd45YNjdPeP4Bx3AVA4\nLZGFM9JZujCbuOhIkxOKBA+PBV1ZWcmOHTtYvXo1AOvXr6e6uhq73c6qVau45557uP/++7FareTm\n5rJixQqvhxYR3zIMg48OnOa5t48AEBMVwb03FLOoJIOUhCiT04kEJ4thGIavd6rTKd6lU1a+EQrj\nbBgGexu62Ph+A922USLCw7h/2QyWLsz2yf5DYYzNpjH2Da+c4haR0DM6Ns5bu1rYdaidrv4RwiwW\ncjLj+es755KlCUVEfEIFLSLnqD/Ry7NvH+FMr50oazhXzc7itiX5TE3TY1IivqSCFhEMw2DPkTPs\nPtRBbWMXFuBrX8lhxXWFREWGmx1PJCSpoEVCmG3Iwa5D7Ww/cJq2ronHKWOiwvnuPQsozk4yOZ1I\naFNBi4SYcZeLg009fHTgFAeauhl3GYSHWbiiNJNr501lXmGqZgQU8QMqaJEQcbp7iO0HTrOzrp3+\noYklHqdnxFNRNpWrZmeREGs1OaGI/CEVtEiQMgyDjl47B5u62XPkDI1t/QDERkVww6JsKuZPIzcr\nXkfLIn5KBS0SRBxj4xw50cfBpm4OHuvmTJ8dAAswpyCVivlTWTgjncgI3fgl4u9U0CIBbsw5zo66\ndvY1dHHkRC9jzokpOKOt4SwqyWB+URrzCtM045dIgFFBiwSo091DVO88zt6jXYw6JtZczs6IY15h\nGvML0yienkREeJjJKUXkYqmgRQKMbcjBG9ub+aD2FC7DICM5mvKF2SyZO4XsDM/LvYpIYFBBiwSI\n0bFxNu1p5e3dLYw4xslKjeWepUUsmJGuG71EgpAKWiQA7Gvo5MXNEwtWxMdE8idfK+K6smk6hS0S\nxFTQIn6ss8/Ok7+q5UyvnfAwC1+/KpdbrsonNlrfuiLBTt/lIn7IMAx2H+7gxfcaGBpxkpIQxbdX\nzKNwWqLZ0UTER1TQIn6mrWvi7uyPD3cQHmah6qaZXFc2lfAwnc4WCSUqaBE/0Tc4yi83NVDT0AlA\ndnocf3XHHN2ZLRKiVNAiJjvRMcAz1Yc52Tnkfu+v7pjDopIM3QQmEsJU0CImMAyDXYfaeab603Pe\nX7m0iBsWZRNt1bemSKjTTwERH+vqt/PG9mZ2HGx3v/dnXy/l2vlT9TyziLipoEV8ZMThZOveNn79\nwTFchsH0jDi+efscpqXHqZhF5DwqaBEvs4862bL3JO9+0sqgfYyYqAiWzJvCndcW6nlmEflC+ukg\n4iX2USeba06y6ZMTDI04iY2K4I5rC6i8Yjqx0ZFmxxMRP6eCFrnMnOMu3tvTylu7WxgacRIXHcGd\nFQUsK8/REbOIfGn6aSFyGR1vt/HsW0doPTNIXHQEK64rZFn5dGKi9K0mIhdGPzVELgPH2DhvbG/m\n3U9acRkGFfOncs8NxcTpVLaIXCQVtMglcI67+MWmBj7cfwqAjORo/nR5KbPzU01OJiKBTgUtcpH2\n1p/hP1/ZT3vPMABL5k7hT742kyhruMnJRCQYqKBFLlB7zzCvftDE7+o7sVhg6cJs7ry2gMQ4q9nR\nRCSIqKBFvqT+wVGeeGkf7T3DGAbMyEnmgWUzyM1KMDuaiAQhFbSIB/ZRJ2/tbmHr3jaGR50A/PWd\nc1l+bSFdXYMmpxORYKWCFvkco45xGtv6+fhwB7sOtTPuMoiNiuD2JfnctiSf8LAwTc8pIl6lghb5\nA6e6htj8u1Z21LUz5nQBkJkSQ/nMDJYvziUhVteZRcQ3VNAiTCz/+MKmBrYfOI1z3EVSnJWry6dQ\nmpvM7PxUrcssIj6ngpaQd7p7iH97oYahkYnry391xxzKZ2YQHqZSFhHzqKAlZDnGxnlzx3He2t0C\nQHZGHFVfm0lJTrLJyUREVNASolwug4c37GB41EmYxcLtS/K5+eo8ncoWEb+hgpaQ9OH+UwyPOimY\nmsjf3j2P5PgosyOJiJxDBS0h57UPj/GbnceJCLfw7RVzVc4i4pdU0BIyRh3jvLi5gY8OnAbg3htm\nkJoYbXIqEZHPp4KWkNDSPsDP3jxER88wuZnxPPC1EmZM181gIuK/VNAS1AzDYNOeVl7Z1sS4y+Cm\nxTncdV0RkRG6GUxE/JsKWoJW48l+Xt9+jMPHe0mMjeQvbp3N3MI0s2OJiHwpKmgJKoZhcPBYD2/t\nbqGhtQ+A4uwkHrptNhnJMSanExH58lTQEjQONHXzv1sbOdU1BMC8wjRuviqXkpxkLWwhIgFHBS0B\nr6vPzi/ea+BAUzcWCyyelcnNV+VpnWYRCWgqaAloA8MO/v2VA7R1DVGam8z9y0qYnhlvdiwRkUum\ngpaA5DIM9tZ38vTrdQAsnJHO39w1T6eyRSRoqKAloAzax9i05wS7D3XQ1T8CwFcXZfPAshKVs4gE\nFRW0BIQe2wi/2NRAbWMXAFGR4VwzdwrXlU3T6lMiEpRU0OLXhkecvL79GDsOnsY+Ok5WaiyLSzO5\n+eo8oiLDzY4nIuI1KmjxW8fbbfzPbz+lrXOIqMhwlsybwoM3z9KpbBEJCSpo8TvDI05+s7OZTXta\nMQy4dt5Uqm4qITJCR8wiEjo8FrRhGKxbt476+nqsViuPP/44OTk57u0HDhzgRz/6EQDp6en8+Mc/\nxmq1ei+xBK1Rxzhv7mzm7d0nAMhMjmHN8pnMzk81OZmIiO95LOjNmzfjcDjYuHEj+/fvZ/369Tz9\n9NPu7WvXruU//uM/yMnJ4ZVXXuHUqVPk5+d7M7MEoQNN3fzXm3XYR8eJi45gTkEqD948C6uuM4tI\niPJY0DU1NVRUVABQVlZGXV2de1tzczPJyck8++yzHD16lKVLl6qc5YKMOsZ55YMm3q85CcCC4nS+\nefscoqwqZhEJbR4LenBwkISE30+ZGBERgcvlIiwsjN7eXmpra3nsscfIycnhm9/8JnPnzuXKK6/0\namgJfIZh8GlLL8+/W8+ZXjtT02J58JZZFE1LMjuaiIhf8FjQ8fHxDA0NuV9/Vs4AycnJ5ObmUlBQ\nAEBFRQV1dXUeCzojQ3Mke5u/jrHLZbB9fxuvbWuk8WQ/FgusWFrMnywvDcjT2f46zsFEY+x9GmP/\n5LGgFy1axNatW1m+fDm1tbWUlJS4t+Xk5DA8PExrays5OTnU1NSwcuVKjzvt7By4tNQyqYyMBL8b\nY5dhsP3Aad7a3cKZXjsA5SUZ3Hx1HgVTE+nvGzY54YXzx3EONhpj79MY+8bF/BLksaArKyvZsWMH\nq1evBmD9+vVUV1djt9tZtWoVjz/+OA8//DAACxcu5Prrr7/gEBLcjrfbeP6deo63T/wQyMmM59sr\n5pKZEmtyMhER/2UxDMPw9U7125p3+ctvxIZh8OLmo+4bwBbPyuSOawuYmhZncrLLw1/GOZhpjL1P\nY+wbXjmCFrlYNfWd7nL+1p1z+UpppsmJREQChwpavGJg2MEv3msA4B/uX8jM3BSTE4mIBBYVtFx2\nzadt/NsLNYy7DOYXpamcRUQuggpaLgvnuIvfHTnD+3tP0tRmA6BwWiJ/t3K+yclERAKTClou2d6G\nTn75XgO9A6NYgLmFqSyZO5UrSjMI08pTIiIXRQUtF80wDN773Uk2vn8UgJLpSVTdNJPsjHiTk4mI\nBD4VtFyUM73DvPrhMT759AwJsZFcNXsK9y2bYXYsEZGgoYKWCzJoH6N653HerznJuMsgJzOev19V\nRkpClNnRRESCigpavrRDzT387I06hkacpCdFs+K6Qr5SmklEeJjZ0UREgo4KWjx67cNj7G/s4sSZ\nQQAqr8hh5dIiIiNUzCIi3qKClkm9/tExfrPzODCxVvNXF2UzrzDN3FAiIiFABS2f69DxHrbtbaOm\noZO46Aj+6o65zClINTuWiEjIUEHLeTZ9coKNWxqBiZWn/vrOuWSlauUpERFfUkGLm8tlcPRkn7uc\nH7p1NlfOydJkIyIiJlBBhzjDMPjowGk++bSD5tM27KPjACTHW7l67hST04mIhC4VdAjrGxzlv944\nRH1rHwBTUmO5clYyZcXpzCvSjWAiImZSQYeo3oFR/vX/7aFv0EF6UjTfvaeMqWlxZscSEZGzVNAh\nZnjEyf/89jD7jnYBE49O/c3d83SdWUTEz6igQ8i2fW08/249ANPS41hcmsnNV+epnEVE/JAKOgTY\nhh28+F4Dn3x6BoC8KQk8uuYKwsJUzCIi/koFHcQMw+Djwx28uPkog/YxiqYl8o2vl2o5SBGRAKCC\nDlL9Qw6efetTDjR1Y40M474bZ3Bj+XQdNYuIBAgVdBBqPtXPd/9jOwCz8lL4xtdLyUiOMTmViIhc\nCBV0kGk82c8TL+1zv/7e6gVYdBOYiEjAUUEHka17T/LCpgYA1iyfyXVl01TOIiIBSgUdBNq6hvj5\nbw5xomNivea7v1rM0gXZJqcSEZFLoYIOYM5xF//5ep170hGAv71rHl9bUkhn54CJyURE5FKpoAOQ\nYRj8+oNjvLW7BYCoyHAeum02C4rTdZe2iEiQUEEHmJr6M7yx/TgnOydOZ8/MSeYvb59DSkKUyclE\nRORyUkEHkOqdx3n1w2MAxERF8A/3LyQ3K8HkVCIi4g0q6ADxmx3NvPZRMwD3fLWYmxbn6A5tEZEg\npoL2cyMOJy9va2Lr3jbSk6J5ZPUCslJizY4lIiJepoL2Y/ZRJ//037vpH3IwLT2Ov1s5n0zNCCYi\nEhJU0H5q1DHOz944RP+Qg9LcZL57TxmREeFmxxIRER9RQfuhzj47z1Qf5ujJflISovjWnXNVziIi\nIUYF7We21bbx4nsNOMcNFhSn88075hAVqXIWEQk1Kmg/8vPfHGLXoQ4AVn21iOWLc3WntohIiFJB\n+wHbkIP/evMQn7b0AvDN2+dw5ewsk1OJiIiZVNAmO3y8h/+7sRYAC/CnXy9VOYuIiAraTM9UH2Zn\nXbv79VMPX0e0Vf9JREREBW0K27CDt3a1uMv5z24u5dp5U3W9WURE3FTQPnaqa4hH/+djDIOJmcHu\nXUBWqmYGExGRc6mgfair386PN+7DMOCqOVl8Y3kpVj1CJSIin0MF7SONbf088eJenOMG18ydwl/c\nOtvsSCIi4sdU0D5QU9/JU68dBKA0N5kHb55lciIREfF3KmgvcrkMnqk+zO7DE5OPrFk+k6ULsk1O\nJSIigUAF7SWGYfC3//4R9lEnAP9cVU5RdpLJqUREJFCEmR0gWG3b1+Yu539SOYuIyAXSEbQX1DV3\n8+Lmo8RERfCvf76Y1MRosyOJiEiA0RH0ZVZ/openXq3DYrHwN3fNUzmLiMhFUUFfRk1t/fzkf/cz\nOjbOt+6Yw6y8FLMjiYhIgNIp7svk07OLXhjALVfnsbAkw+xIIiISwFTQl8GgfYwfn12R6quLsrn7\n+iKTE4mISKDzeIrbMAwee+wxVq9ezZo1a2htbf3cz61du5af/OQnlz2gv3OOu/iX5/YAMCU1lqqv\nzTQ5kYiIBAOPBb1582YcDgcbN27kkUceYf369ed9ZuPGjTQ0NHgloD9zjrv4zk8/oqt/BIBvr5hr\nciIREQkWHgu6pqaGiooKAMrKyqirqztn+759+zh48CCrV6/2TkI/1dDax7pn92AfHQfgyW8vITsj\n3uRUIiISLDwW9ODgIAkJCe7XERERuFwuADo7O9mwYQNr167FMAzvpfQzY85xfvTiXk51DVGam8xT\n372OlIQos2OJiEgQ8XiTWHx8PENDQ+7XLpeLsLCJXn/nnXfo6+vjoYceorOzk9HRUQoLC7nzzjsn\n/ZoZGQmTbvdnwyNj/MvZ9Zxv/EoO37l3IRaLxexY5wnkMQ4kGmfv0xh7n8bYP3ks6EWLFrF161aW\nL19ObW0tJSUl7m1VVVVUVVUB8Nprr9Hc3OyxnAE6OwcuIbK5nvxVLYeae5iVl8KKJfl0dQ2aHek8\nGRkJAT3GgULj7H0aY+/TGPvGxfwS5LGgKysr2bFjh/sa8/r166mursZut7Nq1aoLTxnA3tvTyqHm\nHqKt4Xz3njIiwjXPi4iIeIfHgrZYLPzgBz84572CgoLzPrdixYrLl8oPnegY4KX3jwKw+sYZKmcR\nEfEqtcyX4Bgb57m3jwAwOz+F68qmmZxIRESCnWYS86D1zCBPvXaQM712rpiZwV+vmGd2JBERCQEq\naA9+9Mu9DI86iYoM589unmV2HBERCREq6Em8+mETw6NOAP79767FGhluciIREQkVKugvcPRkH7/d\n1UJinJXv3btA5SwiIj6lm8Q+R/NpG//+8gEMAx68eRbTMzWFp4iI+JaOoP9IQ2sfP/zlXgAeqCxh\nflGayYlERCQU6Qj6j/zyvYlVuXIz47mxfLrJaUREJFSpoP/AOx+foPXMxNSdj37jCpPTiIhIKNMp\nbsAwDPYcOcP/bm0E4M5rCwgP0+8uIiJiHhU08NtdLbz64THCwyzcfm0Bt16dZ3YkEREJcSFf0DX1\nZ3j1w2NEhIfxT1WLyJ+SaHYkERGR0C7oj/af4sXNRwmzWPjOyvkqZxER8RshW9BPvLiXIyf6gIlr\nznMKUk1OJCIi8nsheSfU4eM97nJeVJLB7deev3ymiIiImULuCHrc5eJ/t0zcrf1nXy+lQktHioiI\nHwq5I+gJ1T+iAAANVklEQVRPDp/hxJlBSnOTuXb+VLPjiIiIfK6QK+it+9oAWLowG4vFYnIaERGR\nzxdSBb1tXxuNbf0ALChONzmNiIjIFwuZgh4aGeP5d+sBKJiaqOUjRUTEr4XETWIjDif/9eYhAGbm\nJPMPDywyOZGIiMjkQuII+uVtTdQd6yE7I45HVi8wO46IiIhHQV/QnX12tu6duDHse6sXEhEe9P/K\nIiISBIK6rVwug3/42S4ApqbFkhRnNTmRiIjIlxO0Be0yDP75mY/dr3/w4GIT04iIiFyYoC3oHQdP\n09EzDMDab1yhU9siIhJQgvIu7qa2fp596wjhYRYe/dMryM1KMDuSiIjIBQm6gt5V187Pqw8DcM9X\ni1XOIiISkILuvO/r248BcGP5dJZdMd3kNCIiIhcnqI6gDzX30Nk3AsADlSUmpxEREbl4QXUEXb3z\nOACrlhaZG0REROQSBU1Bn+4eor61D5hYqUpERCSQBU1B/66+E4DlV+YSExVUZ+5FRCQEBUVBu1wG\nmz45AcDSBdNMTiMiInLpgqKgn3/3CEMjTtISo8hMiTU7joiIyCUL+II+3m7jw/2nAfjHB8pNTiMi\nInJ5BHxBP/vWEQAqr8ghLSna5DQiIiKXR0AXtG3YQeuZQQBW6tEqEREJIgFb0C6Xwd//dLv7dWRE\nwP6riIiInCdgW23rvjb3n5/67nUmJhEREbn8ArKgj7fb+OV7DQA8fE+ZnnsWEZGgE5AF/X7NSWDi\ntPbcwjST04iIiFx+AVfQo2Pj7DjYDsCT315ichoRERHvCLiCfvJXtQDkT0kgPibS5DQiIiLeEVAF\nva22jcaT/QD8yddmmpxGRETEewKmoE90DPD8O/UA3HpNHoXTEk1OJCIi4j0BUdCGYfD8uxPlXF6S\nwV3XaVISEREJbgFR0K980MSxUzYyU2L41oq5ZscRERHxOr8vaPuok7d3Tywl+ci9CwizWExOJCIi\n4n1+PcNHj22E7z29E4D0pGgykmNMTiQiIuIbfn0E/dlsYQBrv/EVE5OIiIj4lt8WtMsw3CtV/fQ7\nFXrmWUREQorfFvSnLb109Y9QMX+qyllEREKO3xb0nk87AJiVn2JyEhEREd/zeJOYYRisW7eO+vp6\nrFYrjz/+ODk5Oe7t1dXVPP/880RERFBSUsK6desuOZRhGBw+3gvA3AIthiEiIqHH4xH05s2bcTgc\nbNy4kUceeYT169e7t42OjvLTn/6UX/ziF7z44osMDAywdevWSw7V0Wunq3+EOfkpOr0tIiIhyWNB\n19TUUFFRAUBZWRl1dXXubVarlY0bN2K1WgFwOp1ERUVdcqj3fzexnGRCnPWSv5aIiEgg8niKe3Bw\nkISEhN//AxERuFwuwsLCsFgspKamAvDCCy9gt9u55pprPO40IyPhC7e5XAbv750o6Mor8yf9rHwx\njZtvaJy9T2PsfRpj/+SxoOPj4xkaGnK//qycP2MYBk888QQtLS1s2LDhS+20s3PgC7fV1J8BwBoR\nRkFm3KSflc+XkZGgcfMBjbP3aYy9T2PsGxfzS5DHU9yLFi3igw8+AKC2tpaSkpJztj/66KOMjY3x\n9NNPu091X6wxp4sXzi6K8d17yi7pa4mIiAQyj0fQlZWV7Nixg9WrVwOwfv16qqursdvtzJkzh1df\nfZXy8nKqqqqwWCysWbOGZcuWXVSYXYfasQ2PEREexsxcPV4lIiKhy2NBWywWfvCDH5zzXkFBgfvP\nhw8fvmxhtpy99nzbkvzL9jVFREQCkd9MVGIYBic6Jqb2XL44x8OnRUREgpvfFHR3/wgAs/JSiIwI\nNzmNiIiIufymoLfuawMgJsqvV8AUERHxCb8oaPuok7c/PgHAqq8WmZxGRETEfH5R0Psbu9x/zkqJ\nNTGJiIiIf/CLgv60ZWJhjBXXFZqcRERExD/4RUGf7JyYqWzJ3CkmJxEREfEPphe0c9xF82kb8TGR\npCRc+kIbIiIiwcD0gm5o7QMgKd6KxWIxOY2IiIh/8JuCzkiKMTmJiIiI/zC9oOtPTBT0vTcUm5xE\nRETEf5ha0G9ub6a+tY9p6XFkpOgIWkRE5DOmFvTr25sBuH1JPmG6/iwiIuJmWkGPu1zuPy+elWVW\nDBEREb9kWkHva5iYPUzPPouIiJzPtIL+xXsNABROSzQrgoiIiN8ypaBPdAxgG3IAsHRhthkRRERE\n/JopBf3Z4hilucmanERERORzmFLQtqExACqvyDFj9yIiIn7PlII2MABIT9azzyIiIp/HlIIeHnUC\nYI0wfSIzERERv2RKQx5u7gHAGhluxu5FRET8nikFbQmbuDEsOd5qxu5FRET8ns8L2jAM+gcdxMdE\n6g5uERGRL+Dzgh4ambj+HBGuchYREfkiPi9o2+AoALPyUny9axERkYDh84IeGJ6YQWzAPubrXYuI\niAQMnxd0/+BEQRdnJ/l61yIiIgHD5wXddLIPgDGny8MnRUREQpfPC9ruGAcgJzPe17sWEREJGD4v\n6E+buwEVtIiIyGR8XtA9thEA0pOifb1rERGRgOH7iUrO/n9khKb5FBER+SI+L2in00VinKb4FBER\nmYzPC7p3YJRkFbSIiMikTFkso2/IYcZuRUREAoYpBT01NdaM3YqIiAQMUwq6SLOIiYiITMqUgo6M\nMGW3IiIiAcOUpgwP01KTIiIikzGloA3D8PwhERGREGZKQU9JizNjtyIiIgHDlILOSokxY7ciIiIB\nQ9egRURE/JApBW2xqKBFREQmY1JBm7FXERGRwGFKQYepoUVERCalI2gRERE/pGvQIiIifkhH0CIi\nIn5I16BFRET8kE5xi4iI+CGd4hYREfFDHgvaMAwee+wxVq9ezZo1a2htbT1n+5YtW1i5ciWrV6/m\n5Zdf/lI7VT+LiIhMzmNBb968GYfDwcaNG3nkkUdYv369e5vT6eSHP/whzz33HC+88AK/+tWv6Onp\n8bhTneIWERGZnMeCrqmpoaKiAoCysjLq6urc25qamsjLyyM+Pp7IyEjKy8vZs2ePx52qn0VERCbn\nsaAHBwdJSEhwv46IiMDlcn3utri4OAYGBib9erPyU4mLjrzYvCIiIiEhwtMH4uPjGRoacr92uVyE\nhYW5tw0ODrq3DQ0NkZiYOOnXe+JvKy42q1yAjIwEzx+SS6Zx9j6NsfdpjP2TxyPoRYsW8cEHHwBQ\nW1tLSUmJe1tRUREtLS3YbDYcDgd79uxhwYIF3ksrIiISIiyGYRiTfcAwDNatW0d9fT0A69ev59Ch\nQ9jtdlatWsW2bdvYsGEDhmGwcuVK7rvvPp8EFxERCWYeC1pERER8z5SJSkRERGRyKmgRERE/pIIW\nERHxQypoERERP+S1gvbGHN5yLk9jXF1dzT333MP999/PunXrzAkZ4DyN8WfWrl3LT37yEx+nCw6e\nxvjAgQM88MADPPDAA3znO9/B4XCYlDSweRrnN998k7vuuotVq1bx0ksvmZQyOOzfv5+qqqrz3r/g\n3jO8ZNOmTcY//uM/GoZhGLW1tca3vvUt97axsTGjsrLSGBgYMBwOh3H33Xcb3d3d3ooStCYb45GR\nEaOystIYHR01DMMwHn74YWPLli2m5Axkk43xZ1566SXj3nvvNZ588klfxwsKnsb4jjvuME6cOGEY\nhmG8/PLLRnNzs68jBgVP47xkyRLDZrMZDofDqKysNGw2mxkxA97Pf/5z49ZbbzXuvffec96/mN7z\n2hG0N+bwlnNNNsZWq5WNGzditVqBiYVNoqKiTMkZyCYbY4B9+/Zx8OBBVq9ebUa8oDDZGDc3N5Oc\nnMyzzz5LVVUV/f395Ofnm5Q0sHn6u1xaWkp/fz+jo6OAFjW6WHl5eTz11FPnvX8xvee1gr7cc3jL\n+SYbY4vFQmpqKgAvvPACdruda665xpScgWyyMe7s7GTDhg2sXbsWQ9MJXLTJxri3t5fa2lqqqqp4\n9tln2blzJx9//LFZUQPaZOMMMGPGDO6++25uu+02li5dSnx8vBkxA15lZSXh4eHnvX8xvee1gr7c\nc3jL+SYbY5i45vSjH/2IXbt2sWHDBjMiBrzJxvidd96hr6+Phx56iP/+7/+murqa119/3ayoAWuy\nMU5OTiY3N5eCggIiIiKoqKg478hPvpzJxrm+vp5t27axZcsWtmzZQnd3N++++65ZUYPSxfSe1wpa\nc3h732RjDPDoo48yNjbG008/7T7VLRdmsjGuqqri17/+Nc8//zx/+Zd/ya233sqdd95pVtSANdkY\n5+TkMDw87L6hqaamhuLiYlNyBrrJxjkhIYGYmBisVqv77JvNZjMralD447NqF9N7HlezuliVlZXs\n2LHDfW1u/fr1VFdXu+fw/v73v8+DDz6IYRisWrWKzMxMb0UJWpON8Zw5c3j11VcpLy+nqqoKi8XC\nmjVrWLZsmcmpA4unv8dy6TyN8eOPP87DDz8MwMKFC7n++uvNjBuwPI3zZ098WK1WcnNzWbFihcmJ\nA9tn1/Avpfc0F7eIiIgf0kQlIiIifkgFLSIi4odU0CIiIn5IBS0iIuKHVNAiIiJ+SAUtIiLih1TQ\nIiIifuj/A1YOSWRdCqgMAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11573cb50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Likelihoods:\n",
      "*Increasing the num_sessions by 1 point decreases the chance of label=1 by a factor of 0.9317.*\n",
      "\n",
      "*Increasing the total_artists_followed by 1 point decreases the chance of label=1 by a factor of 0.9892.*\n",
      "\n",
      "*Increasing the total_artworks_favorited by 1 point decreases the chance of label=1 by a factor of 0.9914.*\n",
      "\n",
      "*Increasing the total_artworks_shared by 1 point increases the chance of label=1 by a factor of 1.0498.*\n",
      "\n",
      "*Increasing the gender_male by 1 point decreases the chance of label=1 by a factor of 0.8608.*\n",
      "\n",
      "*Increasing the gender_unknown by 1 point decreases the chance of label=1 by a factor of 0.6957.*\n",
      "\n",
      "To double:\n",
      "*Decreasing the num_sessions score by -9 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the total_artists_followed score by -21 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the total_artworks_favorited score by -69 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the total_artworks_shared score by 10 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the gender_male score by -4 points doubles the chance of label=1.*\n",
      "\n",
      "*Decreasing the gender_unknown score by -1 points doubles the chance of label=1.*\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfEAAAGkCAYAAAAhcRw5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noV7rAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1155c28d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fit_logistic_regression(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 4: Logistic Regression with Label = purchased"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 547,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def preprocess_purchases_and_join_with(df):\n",
    "    # Loading data\n",
    "    purch = pd.read_csv('purchases.csv', usecols=range(1,5))\n",
    "    \n",
    "    # Dropping purchases from anonymous non-registered users\n",
    "    purch.dropna(inplace=True)\n",
    "    \n",
    "    # Aggregating purchase info per user_id\n",
    "    user_purchs = purch.groupby('user_id').agg({'total_pieces_purchased':np.sum, 'total_spent':np.sum})\n",
    "    \n",
    "    # Merging with original DataFrame\n",
    "    user_merged = df.join(user_purchs)\n",
    "    \n",
    "    # Replacing Na's with 0s\n",
    "    user_merged.total_spent.fillna(0, inplace=True)\n",
    "    user_merged.total_pieces_purchased.fillna(0, inplace=True)\n",
    "    \n",
    "    user_merged.drop('total_pieces_purchased', axis=1, inplace=True)\n",
    "\n",
    "    user_merged['purchased'] = np.where(user_merged.total_spent > 0, 1, 0)\n",
    "    \n",
    "    user_merged.drop('total_spent', axis=1, inplace=True)\n",
    "    \n",
    "    return user_merged"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 548,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "\n",
    "# Dropping date columns due to high relationship and used to create the label:\n",
    "df.drop('diff_last_fav_artwork_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_last_artist_folw_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_created_to_last', axis=1, inplace=True)\n",
    "df.drop('days_from_last_session', axis=1, inplace=True)\n",
    "\n",
    "merged_df = preprocess_purchases_and_join_with(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing same model with Statsmodel to see how statistically significant the features are:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sm = SMOTE(kind='regular')\n",
    "X_resampled, y_resampled = sm.fit_transform(X, y)\n",
    "\n",
    "# Splitting train and test data\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3)\n",
    "\n",
    "X_const = add_constant(X_train, prepend=True)\n",
    "\n",
    "logit_model = Logit(y_train, X_const).fit()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Logit Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>         <td>y</td>        <th>  No. Observations:  </th>  <td> 38406</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>               <td>Logit</td>      <th>  Df Residuals:      </th>  <td> 38399</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>               <td>MLE</td>       <th>  Df Model:          </th>  <td>     6</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>          <td>Mon, 27 Jun 2016</td> <th>  Pseudo R-squ.:     </th>  <td>0.1131</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>              <td>15:03:13</td>     <th>  Log-Likelihood:    </th> <td> -23611.</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>converged:</th>           <td>True</td>       <th>  LL-Null:           </th> <td> -26621.</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th> </th>                      <td> </td>        <th>  LLR p-value:       </th>  <td> 0.000</td> \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "    <td></td>       <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th> <td>    0.9967</td> <td>    0.025</td> <td>   40.486</td> <td> 0.000</td> <td>    0.948</td> <td>    1.045</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>x1</th>    <td>   -0.0729</td> <td>    0.001</td> <td>  -48.924</td> <td> 0.000</td> <td>   -0.076</td> <td>   -0.070</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>x2</th>    <td>   -0.0160</td> <td>    0.005</td> <td>   -3.137</td> <td> 0.002</td> <td>   -0.026</td> <td>   -0.006</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>x3</th>    <td>   -0.0093</td> <td>    0.001</td> <td>  -10.063</td> <td> 0.000</td> <td>   -0.011</td> <td>   -0.007</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>x4</th>    <td>    0.0642</td> <td>    0.013</td> <td>    4.765</td> <td> 0.000</td> <td>    0.038</td> <td>    0.091</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>x5</th>    <td>   -0.1583</td> <td>    0.028</td> <td>   -5.649</td> <td> 0.000</td> <td>   -0.213</td> <td>   -0.103</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>x6</th>    <td>   -0.3460</td> <td>    0.028</td> <td>  -12.257</td> <td> 0.000</td> <td>   -0.401</td> <td>   -0.291</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                           Logit Regression Results                           \n",
       "==============================================================================\n",
       "Dep. Variable:                      y   No. Observations:                38406\n",
       "Model:                          Logit   Df Residuals:                    38399\n",
       "Method:                           MLE   Df Model:                            6\n",
       "Date:                Mon, 27 Jun 2016   Pseudo R-squ.:                  0.1131\n",
       "Time:                        15:03:13   Log-Likelihood:                -23611.\n",
       "converged:                       True   LL-Null:                       -26621.\n",
       "                                        LLR p-value:                     0.000\n",
       "==============================================================================\n",
       "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
       "------------------------------------------------------------------------------\n",
       "const          0.9967      0.025     40.486      0.000       0.948       1.045\n",
       "x1            -0.0729      0.001    -48.924      0.000      -0.076      -0.070\n",
       "x2            -0.0160      0.005     -3.137      0.002      -0.026      -0.006\n",
       "x3            -0.0093      0.001    -10.063      0.000      -0.011      -0.007\n",
       "x4             0.0642      0.013      4.765      0.000       0.038       0.091\n",
       "x5            -0.1583      0.028     -5.649      0.000      -0.213      -0.103\n",
       "x6            -0.3460      0.028    -12.257      0.000      -0.401      -0.291\n",
       "==============================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logit_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 549,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Defining y label and X matrix\n",
    "y = merged_df.pop('purchased').values\n",
    "X = merged_df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 550,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({0: 55649, 1: 2167})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 53482 new samples ...\n",
      "done!\n",
      "\n",
      "Logistic Regression Scores:\n",
      "\n",
      "Accuracy on test set: 0.790296495957\n",
      "Precision: 0.853191796649\n",
      "Recall: 0.703944232602\n",
      "\n",
      "K-Fold Cross Validation on Logistic Regression Scores:\n",
      "\n",
      "accuracy: 0.961221807112\n",
      "precision: 0.414696265703\n",
      "recall: 0.0623382873834\n",
      "\n",
      "Beta scores:\n",
      "num_sessions: 0.0049\n",
      "total_artists_followed: 0.0395\n",
      "total_artworks_favorited: 0.0046\n",
      "total_artworks_shared: 0.0286\n",
      "gender_male: 0.4052\n",
      "gender_unknown: 0.8201\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAECCAYAAAAb5qc/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAHdpJREFUeJzt3XmQHGed5vFv3VXdXX2qWt26D0uvZcvIl3wIX4BhWGzA\nhmUXM8NhxoM5lx0TQwyzsTEbxGzArhfPcgwBjInhGAbvjgcPsxwewODBlkHYMrZ1WK8syZKso1ut\nbvVZXXfuH9WXZKm7+qjK6qznE+GwMrOq8tdvSE+9/eabb/ocx0FERLzJ73YBIiJSPgp5EREPU8iL\niHiYQl5ExMMU8iIiHqaQFxHxsJJC3hhzrTHmV+fZ/1ZjzO+MMduNMfcsfHkiIjIfM4a8MebPgL8F\nIufsDwIPALcCtwAfMsYkylCjiIjMUSk9+QPAnefZvwl4yVo7aK3NAk8CNy1kcSIiMj8zhry19hEg\nd55DjcDAlO0hoGmB6hIRkQUwnwuvgxSDflwc6J9fOSIispCCs3it75ztF4GLjDHNQJLiUM39M32I\n4ziOz3fuR4mIlK5QcBhKZugfTnO6f5RCwSGbK5DJ5hlJ5UhncmRyBVLpHKPpHJlsgYPH+1nSHCOb\nK9A7kGI0laWhLkw2VyBfKNDVm8Tvg2AwQCabX5A6w6EAAT/4/X78Ph8Bvw+/H/w+H6cHUixP1BMM\n+AkE/AT8xeOZbIFIOECiOUYw6CcSChCvD/Pef7dpTsE5m5B3AIwxdwH11toHjTH3AT+j+AXwoLX2\n5Ewf4vP56OkZmkutnpNIxNUWY9QWk7zcFoWCQzqbZ3g0y/BoloHhDPmCw6kzScKhAF19STLZPH1D\naeoiQXx+P0MjaZKpHKlMjuRYYKfnGMIvnxg8a/v0QIrGuhCBgJ9Ec5RkKkcsEqQ53kDQ7yOZyrGi\nvYHRdI62pihN9WHqoiGCAR/5vMOS5ih+n4+GWIhQ0E8w4KcuGiQWCeKvks6sz4VVKB2v/gWeLS//\nY54ttcWkxdQWjlPsQQ+PZunpH+X0QIqe/lGy+QKHTw6RTOXoOpMk6Pcxkjrfpb3ShEN+ouEg9dEg\n4VCASNBPMp3jouVN1MdCjKZzLG2tIxT0Ew0FiEWChIL+4nY4SDjkJxwMTPw/GPCx2EYUEol42Xvy\nIuJhBcchmcrRN5hicCRDKpNnNJ1jMJlhYCTDcDLL0GiWV7qHiIQDpDJ5kqkc+cLMHcU00NlWRyZb\nYHminpFUlpXtcUIBPy3xCLFIgHzBYUlTlEDAT2s8Ql00xLKORgb6kwQDum9zrhTyIh7jOA4jY2Gd\nzuY5M5QmPzZmPTKaZSiZZWAkw4Hj/QyOZGmIhUhn84ykspTyi73PB4PJLK2NEVYtjVMXDTKaztHR\nWkdzQ4S6aJC2xigdrXUsbY0RDc89ZuqiIUYU8POikBepco7jMJjM0t2XZHg0y4FjA/j8MDKaI5PL\nc/D4AE0NEQ4eG2Aug6/pbJ7OtmIgx2NhWuIRmhvCxCJBIqEAjfXh4n91YRrqQoSD/kU31FHLFPIi\nLkhn8xO96ly+QC5fYGAkw8neJH3DGXr7kxzpGmIklcPnY8Yedk9/ilgkwGg6T3tzjHhdiGgkSEMs\nxIpEPelsno7WOuqiIeKxEE31YWJVdoFQykMhL1ImhYLDwEiGrr4kx3uGOdlb/P/+YwMzv3mKaDjA\n2s5GYpEgazsbaYlHiNeFaI1HiYYDhEMB6qNB9a7lvBTyIvNQcByO94yw++VeunqTHBqbojcwkmF4\nNHve94SCfprqw6xb1kgmWyDRHCMSDhAK+PD7fZi1bUQDvomLjyLzoZAXmYWRVJbf7z/Ny12D/OrZ\n4xd8XbwuxPpljeQKDp2tdVy6tpXliXqWNMVoiE0f3ItpCqVUP4W8yHlkcwVO9o5weiDFidMjnDg9\ngn2lnzND6YnX+IBYJMiS5iiXrG7lio1LWL6knqjGuaWKKOSlJiVTxWmEh08OsetQL5lcgaFkBseB\nA8fPP2bu88FFK5rYsKKJtsYoV5l2murDFa5cZHYU8uI5BcdhOJnlZO8IR7qHSaayDI5kOD2YYveh\nvhnf39xQDO7N69poqg+zsr2Ble0NLG2pw+9XD10WF4W8LGrDo1l2HezlpeMDHDs1zGAyw+n+FIUL\nzDmMRYLkCwU6WupINMfoaKujqT5MZ1s9KxL1xOvCCnLxFIW8LBrZXJ4Dxwc53jPM0VPDPPnC+dfD\nW7eskeaGCC0NEUJBP6s74ixP1NMajxCLaKqh1BaFvFSlguPQ1ZvkaPcQP/ntETLZAj0Do2fdFBTw\n+4iEAlyzqZ2VS+NsXtvKkqaoQlxkCoW8VIXuviRfefgFjvUUh1wy2cKrXrO2s5H1yxrpbKtj3bIm\nVrTXE/BrXROR6SjkpeIcx2HPy30cOD7AU7u7OD2QOut4wO9j44omAgE/l61r46IVTaxb1qhpiSJz\noJCXsnMch66+JI/uOMqR7iEGhotL10515cXt1IcDvOmaVSxfUu9SpSLeo5CXBec4Dvtf6WcwmWXP\ny338+vkTr3rN1SbBms5Gtm3uoKk+THt7o+7yFCkDhbwsiGyuwO9f6uGnO45ypOvssA4H/XS01nHd\npR10thVv8ddDIEQqQyEvc5bLF3h0x1H2H+vnxcNnznpCUGN9mKtMgi3r29i8tk1zz0VcopCXkh3r\nGaa7L0lXX5KdtofDU3rskVCAmy7vYG1HI1dsXEK9Vk8UqQoKebmgbC7P7kN97Dvaz/5X+jnSffYw\nTFN9mCs3JrhmUzsbVzZrfrpIFVLIy1kGRjI8s+8UT75w8lWhftm6NjatbqGtKcr6ZY20NkZdqlJE\nSqWQr2Gj6Rz/9twJnnjhBKGgn6Pdw2cd72wrru+y9eL2icW6RGRxUcjXmFP9o+zY282z+3teNQum\nIRYiGg5w7SVL2by2FbOqxaUqRWShKORrQDKV45fPHuPnz7zCUHLykXSrl8a5aHkTN27pZGV7g8bU\nRTxIIe9hhYLDr35/nO/9fP/Evkg4wPWXdvDWbWtoiUdcrE5EKkEh70EnTo+wfddJnrGn6Okvrgtz\n1cYEb79xLSsSDS5XJyKVpJD3iL7BFI/8+hDd/aMcODb5+LqbtnTyB9esorNN68GI1CKFvAc88utD\n/L+nDk9sr0g0sG1zBzdu6dRNSSI1TiG/SOXyBZ7a3cW3frpvYt/rr1zOnTetU7CLyASF/CIzOJLh\ndy928w+/eGliX0MsxF1v2MD1mztcrExEqpFCfhEoFBx+s6eLx587zqHjg4wvA7ZxRRN33rRO89lF\n5IIU8lVseDTLN3+0lxcO9U4827SpIcybrl7JJWtaWd0Rd7dAEal6CvkqdGYozV995xnODKUn9m29\nuJ2bL1/GJWtaXaxMRBYbhXwVGU3neGznMX7w60MAtDZGuOGyTrZt7qC9pc7l6kRkMVLIV4HRdI5H\nnjjE478/QS5fwOeDi1e18Kl3X66HV4vIvCjkXfbbPV1892eW0XSehliI11+5ipsvX64lB0RkQcwY\n8sYYH/BVYAuQAu6x1h6acvwPgfuAHPB31tqvlalWzyg4Dk++cJIdL57ixcN9ALzttWv4g2tWEYvo\ne1dEFk4piXIHELHWbjPGXAs8MLZv3P3AJiAJ7DXGfN9aO3CezxEglcnxp1/ZTjqTB2BlewPvuXWD\npkGKSFmUEvI3AI8CWGt3GGOuPuf480ALTEzfdpDzemr3SR567MBEwP/5+7aycZmmQYpI+ZQS8o3A\n1J55zhjjt9YWxrb3ADuBYeAH1trBBa5x0eobTPHNH79I/3CaTLZA72BxRcjL1rXxsTs3s3xZMz09\nQzN8iojI3JUS8oPA1O7mRMAbYy4DbgNWAyPA94wx77TW/tN0H5hIeL/3+uy+U/z1Q8/SPzbXfUlz\njDWdjVx1cTsfuP3SidfVQluUSm0xSW0xSW0xP6WE/HbgduBhY8x1wK4pxwYojsWnrbWOMeYUxaGb\naXm595rNFfjHxw/wi2eOAbBpdQv/+V1bCAX9E68Z//kTibin22I21BaT1BaT1BaT5vplV0rIPwK8\n0RizfWz7bmPMXUC9tfZBY8w3gCeNMWngIPCtOVXiAdlcgU988ddkssWRrA+//VKu2bTU5apEpJb5\nHKfi10kdL34z5/IFPvG/nyCdzRMM+PnCx7YRrwtP+x71UiapLSapLSapLSYlEvE53RmpSdnztPdw\nH9/9V0v3mdGJffd/5PoZA15EpBIU8vPw0GMv8bOnXwFgSVOUtZ2NvOt162lq0N2qIlIdFPJz4DgO\nj+08NhHwV1/czkfv2OxyVSIir6aQn6UzQ2k+9/c7OT2QIhYJ8NE7LuPStVr+V0Sqk0J+FkZSWT71\nN8VJRivbG/hP73wNbU1Rl6sSEbkwhfws/M0PircImJXNfOrdlxMM+Gd4h4iIuxTyMyg4DtlcgYcf\nP8i+o/0E/D7u+49bFPAisigo5Kfx1//3eXYd6j1r3wffsolQMOBSRSIis6OQv4Bn9p2aCPjN61qJ\nhoNcf+lSrtiQcLkyEZHSKeTPI5nK8Z1/tQC85brV/Ptb1rtckYjI3Cjkz5HLF/jkl54gX3C4YsMS\n3nnzOrdLEhGZM109PMeDP9pLvuAQCQX4yB2b8elB2iKyiCnkpxgcyfC7F08B8IWPbdMMGhFZ9JRi\nYwqOw9f/ZQ8AV2xYQl005HJFIiLzp5Af87Uf7uHFI2dY29nIh9+udWhExBsU8sDhrkGe2Vccprn3\nbZec9RQnEZHFrObTbCSV5bPfegYoTpdsb6lzuSIRkYVT8yH/v77/HABLW+t4h6ZLiojH1HTID49m\nOdo9RCwS4NN3XYFf0yVFxGNqOuSf2t2FA7x2cyctcT3NSUS8p2ZDfiSV5aHHXgLgio1aj0ZEvKlm\nQ/4r/1RcG35JU5RNq1tcrkZEpDxqMuSTqRz2lX4A/st7r3K5GhGR8qnJkH9sZ/EB3FduTNDUoLF4\nEfGumgv5dDbPYzuPEQ75ee+bNrpdjohIWdVcyP/oqcMMJrO8aesq9eJFxPNqKuS7+pL8+DdHCPh9\nvGnrSrfLEREpu5oK+d1jj/O7fnMHDTGtMiki3ldTIX+0exiAaza1u1yJiEhl1EzIDyUzPLnrJIDm\nxYtIzaiZkP/nJ18Gir34gL9mfmwRqXE1kXan+0d54vmTxOtCvP/NF7tdjohIxXg+5B3H4XPfe5Zc\nvsC7X7+BWCTodkkiIhXj+ZDfsbebM0NpouEA11661O1yREQqqiZCHuB9bzZaL15Eao6nQ/7MUJrn\nDxbnxl9tNG1SRGrPjAPUxhgf8FVgC5AC7rHWHppyfCvwhbHNLuCPrLWZMtQ6a8+91APAxpXNBAOe\n/j4TETmvUpLvDiBird0GfAZ44Jzj3wA+YK29CXgUWL2wJc7dM7YY8h966yUuVyIi4o5SQv4GiuGN\ntXYHcPX4AWPMRqAXuM8Y8zjQaq19qQx1ztpQMoM92s+6ZY20NkbdLkdExBWlhHwjMDBlO2eMGX/f\nEuB64EvArcCtxphbFrTCOXAchy889BwFx+Eqo0f7iUjtKiXkB4H41PdYawtjf+4FDlhr91trcxR7\n/Fef+wGV9sQLJzl6qrhOzc1blrlcjYiIe0q5M2g7cDvwsDHmOmDXlGOHgAZjzLqxi7E3Ag/O9IGJ\nRHyml8xZJpvn//yyOGL0Vx/exuqVrWU710IoZ1ssNmqLSWqLSWqL+Skl5B8B3miM2T62fbcx5i6g\n3lr7oDHmj4HvG2MAnrLW/nSmD+zpGZpzwTP5l+0vM5rO094cY1lztKznmq9EIl7V9VWS2mKS2mKS\n2mLSXL/sZgx5a60DfOSc3funHH8cuHZOZy+D8eWE3/W69S5XIiLiPs9NHj9xeoRwyM+VG3XBVUTE\nUyE/ms7R1Zdk/bImfFrCQETEWyF/pKs4dre6QxdqRETAYyG/53AfAGsU8iIigIdCfsfebn78myMA\nrO1sdLkaEZHqsOifoFEoOPz9z/fz+O+PA3DTlmUkmmMuVyUiUh0Wfch//nvPcuB4cdWF265fzTtv\n1tRJEZFxizbkc/kC3/7pPg4cH6AlHuFjd17GumUaphERmWpRhnzfYIr//t2dnBlK0xKP8On3XMHS\nljq3yxIRqTqLLuTzhQL3P/QcZ4bSXH7REj542yYaYiG3yxIRqUqLLuS/9s976O5LsmFFEx9/x2X4\n/brpSUTkQhbVFMpXTg2zc3/xaU9/fNsmBbyIyAwWTU/+SNcQn/veTgDeePVK2jUGLyIyo0UT8t9+\ndB+ZbIE3X7uK//C6i9wuR0RkUVgUwzXDo1kOdw0Rrwsp4EVEZqHqQ95xHL74j88DsE7LFYiIzErV\nh/zT+05x8MQgq5fGufftl7pdjojIolL1If+DfzsEwBuuWkE0vGguIYiIVIWqDvmf/e4op/pHCfh9\nXLlxidvliIgsOlUb8tlcgYd+eQCAT7/nCuqiuqtVRGS2qjbk7dEzAFyxYQkbVjS7XI2IyOJUtSG/\nY283AK+7crnLlYiILF5VGfIvnxxk++4umurDXLK61e1yREQWraoM+V89W3zK05uvXaX1aURE5qEq\nQ34wmQHgtZd1ulyJiMjiVnUhn80VeOFgL8GAj7qI5sWLiMxH1YX8N3+8F4DmhoiGakRE5qnqQn7P\ny30A3Ps2LWEgIjJfVRfyPl+x975+eZPLlYiILH5VFfKO4zA8mmVNR9ztUkREPKGqQv7giUEAouGA\ny5WIiHhDVYX8z55+BYDLNyRcrkRExBuqKuSf2XcKgBs0P15EZEFUTcgXHIfxCZN1Uc2PFxFZCFUT\n8slUDofiqpMiIrIwqibku/uSANRr3XgRkQVTNSH/9Nh4vOM4LlciIuIdMw5+G2N8wFeBLUAKuMda\ne+g8r/s60Gut/Yu5FGKP9gNw3eaOubxdRETOo5Se/B1AxFq7DfgM8MC5LzDG3AtsnmsRjuNwpHsI\ngEvXaP14EZGFUkrI3wA8CmCt3QFcPfWgMeZ6YCvw9bkWkc7mAWhqCM/1I0RE5DxKCflGYGDKds4Y\n4wcwxnQAfwl8HJjzkpG5fHEcfv0yrVcjIrKQSpmQPghMXUzGb60tjP35XUAb8BOgE4gZY/ZZa78z\nmyLy+eLHBQNaWlhEZCGVEvLbgduBh40x1wG7xg9Ya78MfBnAGPN+wJQS8InE2QuQpca+Murrwq86\n5nW19vNOR20xSW0xSW0xP6WE/CPAG40x28e27zbG3AXUW2sfnMtJe3qGztp+evcJAEaSmVcd87JE\nIl5TP+901BaT1BaT1BaT5vplN2PIW2sd4CPn7N5/ntd9e04VAC8ePgPA1ouXzvUjRETkPKriZqjf\n7u0mHPJz+YY2t0sREfEU10M+nSlOn1zSFCPgd70cERFPcT1Vs2Mza5a2xFyuRETEe1wP+UKhOEc+\nEHC9FBERz3E9WfPjIe/XHHkRkYXmesiP9+T9PoW8iMhCcz3k84568iIi5eJ6yGfGFifzK+RFRBac\n6yE/vo58fUzPdRURWWiuh/y+I8W7XSPBgMuViIh4j+shP75AsZ4IJSKy8FwP+fHZNfVRDdeIiCy0\nqgl5TaEUEVl47od8MeMV8iIiZVAFIT/Wk3e9EhER73E9WieGazRPXkRkwVVPyGu4RkRkwbke8qcH\nRgHwKeRFRBac6yHfO5h2uwQREc9yNeRHUlkAlrbWuVmGiIhnuRryuw/1AXDxqmY3yxAR8SxXQ34w\nmQFgbWejm2WIiHiWqyGfG3u+a2N92M0yREQ8y+WQL06fDAY0s0ZEpBzcDflcsScf0kO8RUTKwt2Q\nLxRDPqCQFxEpC5d78sXhGvXkRUTKo0p68hqTFxEpB43Ji4h4mKvp2jeYAtSTFxEpF1dD/kRvEoB4\nnebJi4iUg2shPzCS4cxQmotWNBEJBdwqQ0TE01wL+fElhtd0xN0qQUTE81wL+fGLrrFw0K0SREQ8\nz7WQz46tWxMKamaNiEi5uBfyYz35oKZPioiUjXvDNWOLk6knLyJSPjMOiBtjfMBXgS1ACrjHWnto\nyvG7gE8CWWCXtfajpZw4m8sDCnkRkXIqJWHvACLW2m3AZ4AHxg8YY6LAZ4GbrbU3As3GmNtLOXFW\nd7uKiJRdKQl7A/AogLV2B3D1lGNpYJu1dvxp3EGKvf0ZTawlr568iEjZlJKwjcDAlO2cMcYPYK11\nrLU9AMaYTwD11tpflHLirr7i3a7qyYuIlE8pk9QHgal3LPmttYXxjbEx+/8JbADeUcpJE4k4qbHh\nmo6lcRKJ2r0hqpZ/9nOpLSapLSapLeanlJDfDtwOPGyMuQ7Ydc7xbwCj1to7Sj1pT88Qv911EoC6\noI+enqFS3+opiUS8Zn/2c6ktJqktJqktJs31y66UkH8EeKMxZvvY9t1jM2rqgZ3A3cATxphfAQ7w\nRWvtD6f7QMdxyBcc6qNBGrU4mYhI2cwY8tZaB/jIObv3z+YzzjU0mgUg0Ryb7VtFRGQWXLnqefhk\n8dev1VqcTESkrFwJ+b2H+wBojUfcOL2ISM1wJeSf3ncKgA0rmt04vYhIzah4yDuOQyqTA8CsUsiL\niJRTxUN+JJVjNJ3nNevb8Pn0bFcRkXKqeMj3nCne6drWGK30qUVEak7FQ75/qLjMTVOD5seLiJRb\nxUN+eGyOvB7eLSJSfhUP+e6xhckcp9JnFhGpPRUP+UPHiwtarlraUOlTi4jUnIqH/MneEUB3u4qI\nVELFQ358/fj6aKjSpxYRqTkVD/mXTwwQDulBISIilVDxtG2IhchkCzO/UERE5q3iIT80mmVFQhdd\nRUQqoeIhn87kqYtojryISCW4Mjge0MO7RUQqwqWQ18JkIiKV4ErIB/3qyYuIVII7IR9UyIuIVIIr\naasLryIileFKyPv1sBARkYpwJeR9foW8iEgluNOTRyEvIlIJ7vTklfEiIhXhSsjnC3piiIhIJWgu\no4iIh7kS8m1NUTdOKyJSc9y5GUpr14iIVIRLIa8rryIilaCevIiIh6knLyLiYerJi4h4mCtpWx8N\nuXFaEZGaE6z0Cf/bn1zHipZYpU8rIlKTKt6Tv+ripfi1QJmISEXM2JM3xviArwJbgBRwj7X20JTj\nbwX+K5AF/s5a+2CZahURkVkqpSd/BxCx1m4DPgM8MH7AGBMc274VuAX4kDEmUYY6RURkDkoJ+RuA\nRwGstTuAq6cc2wS8ZK0dtNZmgSeBmxa8ShERmZNSQr4RGJiynTPG+C9wbAhoWqDaRERknkoJ+UEg\nPvU91trClGONU47Fgf4Fqk1EROaplCmU24HbgYeNMdcBu6YcexG4yBjTDCQpDtXcP8Pn+RKJ+Awv\nqR1qi0lqi0lqi0lqi/nxOc70D/CYMrvmNWO77gauAuqttQ8aY24D/hLwAd+01n6tjPWKiMgszBjy\nIiKyeGkRGRERD1PIi4h4mEJeRMTDFPIiIh5WtlUotebNpBLa4i7gkxTbYpe19qOuFFpmM7XDlNd9\nHei11v5FhUusmBL+TmwFvjC22QX8kbU2U/FCK6CEtvhD4D4gRzErPD+DzxhzLfB5a+3rztk/69ws\nZ09ea95Mmq4tosBngZuttTcCzcaY290ps+wu2A7jjDH3ApsrXZgLZmqLbwAfsNbeRHFZkdUVrq+S\nZmqL+4HXU1xi5VPGGE/fVW+M+TPgb4HIOfvnlJvlDHmteTNpurZIA9ustemx7SDF3owXTdcOGGOu\nB7YCX698aRV3wbYwxmwEeoH7jDGPA63W2pfcKLJCpv17ATwPtADjD6Lw+rzvA8Cd59k/p9wsZ8hr\nzZtJF2wLa61jre0BMMZ8guJNZr9wocZKuGA7GGM6KN5U93GKN9Z53XT/PpYA1wNfothru9UYc0tl\ny6uo6doCYA+wk+Ld9j+y1g5WsrhKs9Y+QnFo6lxzys1yhrzWvJk0XVtgjPEZY+4H3gC8o9LFVdB0\n7fAuoA34CfDnwHuMMe+rcH2VNF1b9AIHrLX7rbU5ir3cc3u3XnLBtjDGXAbcRnG4ag2w1BjzzopX\nWB3mlJvlDPntwFsAplvzxhgTpvgrx2/KWIvbpmsLKI6/Rqy1d0wZtvGiC7aDtfbL1tqt1trXA58H\n/sFa+x13yqyI6f5OHAIajDHrxrZvpNib9arp2mKA4rpYaWutA5yiOHRTC879jXZOuVm2ZQ205s2k\n6dqC4q+hTwNPjB1zgC9aa39Y6TrLbaa/E1Ne937A1Mjsmgv9+7gF+B9jx56y1v5p5ausjBLa4l7g\ngxSvXx0E/mTsNxzPMsasBr5vrd02NvtuzrmptWtERDxMN0OJiHiYQl5ExMMU8iIiHqaQFxHxMIW8\niIiHKeRFRDxMIS8i4mEKeRERD/v/O0W1nuFvOwoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1354ab8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Likelihoods:\n",
      "*Increasing the num_sessions by 1 point increases the chance of label=1 by a factor of 1.0414.*\n",
      "\n",
      "*Increasing the total_artists_followed by 1 point increases the chance of label=1 by a factor of 1.1143.*\n",
      "\n",
      "*Increasing the total_artworks_favorited by 1 point increases the chance of label=1 by a factor of 1.0095.*\n",
      "\n",
      "*Increasing the total_artworks_shared by 1 point increases the chance of label=1 by a factor of 1.0865.*\n",
      "\n",
      "*Increasing the gender_male by 1 point increases the chance of label=1 by a factor of 1.4752.*\n",
      "\n",
      "*Increasing the gender_unknown by 1 point increases the chance of label=1 by a factor of 2.7088.*\n",
      "\n",
      "To double:\n",
      "*Increasing the num_sessions score by 142 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the total_artists_followed score by 17 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the total_artworks_favorited score by 149 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the total_artworks_shared score by 24 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the gender_male score by 1 points doubles the chance of label=1.*\n",
      "\n",
      "*Increasing the gender_unknown score by 0 points doubles the chance of label=1.*\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfcAAAGkCAYAAAAsb2x+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npMwY7pIkZcZwlyQpM/8LugQWUWEo6h8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x127d15f10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fit_logistic_regression(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Confusion matrix notes:\n",
    "\n",
    "##### We want to minimize False Positives (upper right corner), because if we predict that they will make a purchase and they don't is where we spent the most money/effort without results. \n",
    "##### In case of a False Negative (lower left corner), we didn't predict purchase but they bought it anyways, \"free\" profits $$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 5: Random Forests with Label = purchased (with best params after GridSearch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 767,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "plt.rcParams.update({'font.size': 22})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 897,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def fit_random_forest(X, y):\n",
    "    sm = SMOTE(kind='regular', ratio=0.5)\n",
    "    X_resampled, y_resampled = sm.fit_transform(X, y)\n",
    "    \n",
    "    # Splitting train and test data\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3)\n",
    "    \n",
    "    rf = RandomForestClassifier(oob_score=True, n_jobs=-1, bootstrap=True, min_samples_leaf=2, \n",
    "                                n_estimators=400, min_samples_split=1, random_state=67, \n",
    "                                max_features=None, max_depth=None)\n",
    "    rf.fit(X_train, y_train)\n",
    "    \n",
    "    # Draw a confusion matrix for the results\n",
    "    y_predict = rf.predict(X_test)\n",
    "    y_proba = rf.predict_proba(X_test)\n",
    "    cm = standard_confusion_matrix(y_test, y_predict)\n",
    "    \n",
    "    print \"Random Forest Scores:\\n\"\n",
    "    print \"accuracy:\", rf.score(X_test, y_test)\n",
    "    print \"precision:\", precision_score(y_test, y_predict)\n",
    "    print \"recall:\", recall_score(y_test, y_predict)\n",
    "    \n",
    "    tpr, fpr, thres = roc_curve(y_proba[:,0:1].flatten(), y_test)\n",
    "    plt.plot(tpr, fpr)\n",
    "    plt.title(\"\\nROC Curve\\n\", fontsize='small')\n",
    "    plt.xlabel(\"False Positive Rate (1-Specificity)\", fontsize=13)\n",
    "    plt.ylabel(\"True Positive Rate (Sensitivity)\", fontsize=13)\n",
    "    plt.savefig('images/roc_curve4.png', bbox_inches='tight', pad_inches=0.5)\n",
    "    plt.show()\n",
    "    \n",
    "    fix, ax = plt.subplots(figsize=(10, 7))\n",
    "    sns.heatmap(cm, annot=True,  fmt='', square=True, \\\n",
    "                            xticklabels=['1', '0'], \\\n",
    "                            yticklabels=['1', '0']);\n",
    "    plt.title(\"\\nConfusion Matrix - Random Forests\\n\", fontsize='small')\n",
    "    plt.xlabel(\"Truth\", fontsize=13)\n",
    "    plt.ylabel(\"Prediction\", fontsize=13)\n",
    "    plt.savefig('images/confusion_matrix4.png', bbox_inches='tight', pad_inches=0.5)\n",
    "    plt.show()\n",
    "\n",
    "    cols = list(df.columns)\n",
    "    \n",
    "    print \"Feature Importance: \\n\"\n",
    "    for name, importance in izip(cols, rf.feature_importances_):\n",
    "        print round(importance,4), '\\t\\t', name\n",
    "        \n",
    "    plot_importance(rf, merged_df, max_features=16)\n",
    "    plt.savefig('images/rf_feature_importance4.png', bbox_inches='tight', pad_inches=0.5)\n",
    "    \n",
    "    return rf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 898,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "df = df[~df.index.duplicated(keep='first')]\n",
    "\n",
    "# Dropping date columns due to high relationship and used to create the label:\n",
    "df.drop('diff_last_fav_artwork_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_last_artist_folw_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_created_to_last', axis=1, inplace=True)\n",
    "df.drop('days_from_last_session', axis=1, inplace=True)\n",
    "\n",
    "merged_df = preprocess_purchases_and_join_with(df)\n",
    "merged_df = merged_df[~merged_df.index.duplicated(keep='first')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 899,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Scaling features\n",
    "merged_df.num_sessions = scale(merged_df.num_sessions)\n",
    "merged_df.total_artists_followed = scale(merged_df.total_artists_followed)\n",
    "merged_df.total_artworks_favorited = scale(merged_df.total_artworks_favorited)\n",
    "merged_df.total_artworks_shared = scale(merged_df.total_artworks_shared)\n",
    "\n",
    "# Defining y label and X matrix\n",
    "y = merged_df.pop('purchased').values\n",
    "X = merged_df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 900,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({0: 53167, 1: 2074})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 24509 new samples ...\n",
      "done!\n",
      "Random Forest Scores:\n",
      "\n",
      "accuracy: 0.954775339603\n",
      "precision: 0.94963744232\n",
      "recall: 0.911426040744\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAFHCAYAAABprdEfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x139ecfbd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhoAAAHuCAYAAADZUqW0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x15d7e88d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature Importance: \n",
      "\n",
      "0.7979 \t\tnum_sessions\n",
      "0.0402 \t\ttotal_artists_followed\n",
      "0.1298 \t\ttotal_artworks_favorited\n",
      "0.0112 \t\ttotal_artworks_shared\n",
      "0.0075 \t\tgender_male\n",
      "0.0134 \t\tgender_unknown\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAg8AAAFHCAYAAAAmxq8+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhAcAoEV4AABahAcAoEV4AABafgB0j+sdU3AqLwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x147298290>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = fit_random_forest(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 742,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 742,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 788,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.3075905 ,  1.76841926,  0.03038993,  0.56564144,  1.        ,  0.        ])"
      ]
     },
     "execution_count": 788,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 747,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'numpy.ndarray' object has no attribute 'hist'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-747-d846c034ab9c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m: 'numpy.ndarray' object has no attribute 'hist'"
     ]
    }
   ],
   "source": [
    "model.predict_proba(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 754,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({0: 55649, 1: 2167})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 25657 new samples ...\n",
      "done!\n",
      "Random Forest Scores:\n",
      "\n",
      "accuracy: 0.951122114847\n",
      "precision: 0.942324317649\n",
      "recall: 0.909741266245\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEZCAYAAACNebLAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XecXVW5//HPmcm0JDOpQxIiNcATioYqSaS3e1VAilyN\nqAiiiGBDfSnoFUGv9SeCKApEEY0NQS5VLEiRiFwIxQDyQEjobdInZcqZc35/rH1yDsMpeybZZ9r3\n/XrlldnlrPOclcl+zlp7r7VS2WwWERGR3moGOgARERmclCBERKQoJQgRESlKCUJERIpSghARkaKU\nIEREpKhRAx2AyJZmZhlgMZABssBoYA3wcXdfFJ0zGrgAOAbojM67Cfgfd+8oKOsU4AygEagH7gG+\n4O5rSrx3n84XGczUgpDhKAsc4u57ufve7j4TuAa4FMDMaoG/AilgT3efBcwGmoE/mVlNdN55wGnA\nse6+NzALSAM3FnvTvp4vMtilNFBOhpuoBTHZ3VdG27XARcCO7n6Mmc0DPunuc4q89iHg68AfgVeB\nWe6+tOB4I3A88Ht3TxfsH13h/GuB84BJ7v7J6Nj5uW0zuwNYCRhwBfDfwDR3T0cJ61ngSOAl4BJg\nD6AOuB34vLtnNrfeRHpTF5MMV3eYWRZoBToI3UenRsfmAHeXeN3twAHAMmB94cUeIOp++k2R182s\ndL6ZVYp5pbvvEZ17PHAs8AfgP4Bl7v6Emf0UeMDdT40Sx8+BzwLfrVS4SF+pi0mGq0PcfU/gnUAT\n8A93X15wvK7E6xoIXVQZ+vb/o6/nF/P3gp/nAx+Kfv4QcGX089HAGVFLZxGwH6E1IbLFKUHIcJUC\ncPeHgXOAn5rZttGxhcBBvV9gZqlo/0LgcaDOzHbsdU6Dmd1iZlN7vTzO+dlcXJH6XmWsK/j5WmB/\nM5sZxXRNtL8GOCm6v7IXsD9wdqlKENkcShAy7Ln7b4F/EPruIVx815vZxdE9AsysiXATux34X3fv\nAr4N/MzMtorOaQAuBka7+yu93qPc+U3R+W3APtGxMcBRZWLuBH5H6EK6LtoG+BMh4eXKvwklCEmI\nEoQMR8WevPgE8J9mdqS79xAuzuuBRWb2L+ABQnLIHcfdvwVcR3iy6UHgoajsdxV70zLnHxed8iug\nzcyeBG4mJK1yMV9J6EK6smDfp4DRZrYYeBh4BPhOmboQ6Tc9xSQiIkUl3oIws/2jR/h67z/GzP7P\nzBaa2elJxyEiIn2TaIIws88TmscNvfaPIjyXfgRwCPBRM2tNMhYREembpFsQSwiDhHrbFXjK3de6\nezdhOoI3PFUiIiIDJ9EE4e7XE6Ya6K2FMDdOTjswLslYRESkbwZqJPVaQpLIaQZWV3pRNpvNplKp\nSqcNWT2ZLO3ru1i+ZiOvrFjPho40y14MebSjq4dV7R088ewqJo9rZENnmvb1XXR09SQWz6jaGmpr\nU9TWhD81NSnaN3TTWF/L+LEN1ET7alLh79qaFCvWbGTiuCYa62tJkSL3z5VKQSqVIkX4mxTURAfD\nZio6JzpesP+VlevZdkozqVR4r1SKTe+bispJ1eR/blu1ka1bx/Tps5b6vSr521biQKr0Kyj1q1v6\nPUrE1OdySoZUMt6+/jfra/2VLr8f9dfHsrZk/Q0Vxx44o1+foloJondw/wZ2MrPxwAZC91LFqQJS\nqRRtbe0JhJe8TDbLmnVdrFnfyar2Tp5/dR3rNnbzxHOr6OzuoSeTZeXazsoFAes3dtM6oYnW8U00\n1dfS3ZNhyoTRjG2qo7GhlrraGupG1dJQV0N9XS21Nalwsa9J0d2ToXl0fbjo16YYVVNDU+MoGutq\nN23nksJQScatrc1D9vdiS1Nd5KkuNl+1EkQWIJokbYy7zzezc4A/E5LHfHd/uUqxJCrdk2HJC2u4\n46EXGd04iuVrOli+eiOvrtpY8bU7v2kcHV09bD15DKMbRrHVhCamThxNQ10t45sbaBldT2NDLTWp\nlH75RSRxQ20cRHYwXhSffaWdBX92VqztYPW6rjccH9tUx6SWRupG1bDNVmOZPL6RCWMbGN1Yx6SW\nBiaNa6Sxvm+5WgkiT3WRp7rIU13ktbY2D+oupmFlVXsnt/7zWR5/ZiVd3RlWrN20vgxjm+qYud0E\ndpo+jhlbt/Cm1rE01NcOYLQiIv2jBBHTuo3dPLJkOfc9/iqPLlv5umN77jSZt+w0iTm7T6WhTslA\nRIYHJYgyurp7eOK51dzzr5d4wNs27Z+xdQtz9pjKDtNa2HbKWGprNKWViAw/ShAl3P3IS/z8j09s\n2p48rpE9dpjI2948jRnTNWRDRIY/JYgivrlgEU+9EMYf7Lh1CycctCMzt5uw6bl9EZGRQAmiQDab\n5eu/eIBlL4cnHz5yzG7M2b33ujAiIiODEkRkfUc3n7g4v+Lj+4/aRclBREY0JQhCy+Hzl+XXbvnK\nh/Zl+6ktZV4hIjL8KUEAN9yzjI6uHqZMHM35H9q3z4PWRESGoxH/fObti17gxoXP0NRQy6dPeouS\ng4hIZEQniAefbONXf3kSgDPftQdTJowe4IhERAaPEZsg1q7v4od/WAzAgW+Zxh47ThrgiEREBpcR\nmSCy2SwLopbDpJYGTn3HrgMckYjI4DMiO9z//ewqHnjiNbaf2syXPrjPQIcjIjIojcgWxMNLlgPw\n1l2naB4lEZESRtzVMd2T4c6HXgJgzu5TBjgaEZHBa8QliL89+CLpngwztx3PuLENAx2OiMigNeIS\nxMNPhWm7TzhoxgBHIiIyuI2oBPH0i2t44rnVTGppYKc3acpuEZFyRlSCuOXeZwE44WC1HkREKkn0\nMVczSwGXAbOADuB0d19acPwDwOeA1cDV7v6zpGJ57tV2Hl6ynCkTmpi9m25Oi4hUknQL4jigwd3n\nAucCF+UOmNkk4ELgIOAQ4GQz2zapQHJLhm6z1VhSWvhHRKSipBPEAcBtAO5+H7BvwbEdgYfdfY27\nZ4H7gdlJBLGqvZM/3/8cYxpHccrbZybxFiIiw07SI6lbgDUF22kzq3H3DPAUsLuZtQLrgcMBr1Rg\na2tzn4P4/d1L6erO8F9H7ML220zs8+sHq/7UxXClushTXeSpLjZP0gliLVD4L5RLDrj7ajM7B7gO\nWAEsApZXKrCtrb1PAazv6OaP/3gGgAN3n9Ln1w9Wra3Nw+azbC7VRZ7qIk91kdffRJl0F9NC4B0A\nZjYbWJw7YGa1wN7ufhDwHmBmdP4W9cSzqwCYMb2FulG1W7p4EZFhK+kWxPXAkWaWu/CfambzgDHu\nPt/MMLMHgY3A99x95ZYO4M6HXgTgA0fZli5aRGRYSzRBRDefz+y1+8mC4xcSnmRKxLKX1/LYM6uw\nbcaz7RT1RYqI9MWwHih308JnAJizx9SBDUREZAgatgmiJ5Ph+dfCDar9Zm41wNGIiAw9wzZBPLp0\nJSvWdnLIXtNpahiR6yKJiGyWYZsgFi5+GQjrTYuISN8NywSxbmM3Dy9ZzvTJY9h+qm5Oi4j0R6y+\nFzN7M7AzkAGWuPujiUa1me57/FXSPVne9uZpmndJRKSfSiaIaCbWjwGfBtqB54BuYAczawEuAS7P\njYweTP719AoAZmtJURGRfivXgrgW+Asw291XFR4ws3HAKYSBcO9KLrz+eWXlelrG1DNeS4qKiPRb\nuQTxQXdfX+yAu68BfmBmP00mrP7rTmdYvqaDnadrxTgRkc1RMkHkkoOZPQpcDfzS3V8pds5g8vKK\n9WSzMGXi6IEORURkSIvzFNM7gUbgDjO7xczebWZ1CcfVb7nJ+RrrNfZBRGRzVEwQ7v6su3/N3XcF\n5gPfB142s4ujVeEGldXruwCYNlktCBGRzVHxa7aZjQXeDXwAmA78GPgd8B/An3j9KnEDLvcE0zZb\njR3gSEREhrY4/TDLgJuBC9z97txOM/sxcGRSgfXXS8vDbZEdprUMcCQiIkNbnATxYXe/sXCHmZ3g\n7n8Ajk8mrP7pTueHZNRogJyIyGYpN1DuPUADcKGZjS84VAecC/wh4dj67OUVofVg24yvcKaIiFRS\nrgXRAswlrCl9aMH+NPClJIPqr0eWhCWt999NI6hFRDZXuXEQVwJXmtnh7n57FWPqt4eXLKe2JsV+\nu2r9BxGRzVWui+kKd/8o8GUze0OLwd0PSzSyPlrV3smyl9vZdbsJjGkctMM0RESGjHJdTJdHf3+1\nv4VHE/5dBswCOoDT3X1pwfGTgXMI3VZXuftP+vtejy1bCcCsGYNuaIaIyJBUrotpUfTjOcAvgRvd\nvauP5R8HNLj7XDPbH7go2pfzXWBXYAPwuJn9Jprnqc8W+WsAbD15TH9eLiIivcSZauMKwkX9aTOb\nb2aH9KH8A4DbANz9Pt44qO4RYALQFG1n+1D26zzftg6A7TX+QURki4gz1cYt7v5+YBfCxf57ZvZs\nzPJbgMIWQdrMCt/zMWARsBi42d3Xxiy3pLFNuv8gIrIlxF1RbjfgvcBJwPPAxTHLX0t4TDanJrfA\nULRK3TuB7YD1wK/M7ER3v65cga2tb1xCNJvNsrEzzYTmhqLHh6uR9FkrUV3kqS7yVBebJ85cTIsJ\nN5EXAIe5+8t9KH8hcDRwrZnNJrQUctYQ7j10unvWzF4jdDeV1dbW/oZ9a9Z3sbGzh5nbTih6fDhq\nbW0eMZ+1EtVFnuoiT3WR199EGacF8T53X1z5tKKuB440s4XR9qlmNg8Y4+7zzewK4B4z6wSeBn7e\nnzfJzb+kG9QiIltOnHEQPzCzN9w8jjMOwt2zwJm9dj9ZcPxy8o/T9lsuQUxXghAR2WISHQdRLS9G\nTzBNb9UU3yIiW0qccRDvdvdPFB4zs6uBu5IMrC9eWL6emlSKqVpmVERkiynXxTQf2BHY18x27/Wa\nQTNdajab5cW29UyZ2ETdqDjDOkREJI5yXUxfB7YHLgEuKNifBv6dYEx9snpdFxs70+y2fcUHoERE\npA/KfeXucPc7gWMIq8rl/jwPDJrO/k33H3SDWkRkiyrXgphPGMNwF2EKjMIl2rKE7qcB92g0SZ/u\nP4iIbFnlblIfHf29Q/XC6bu21RsB2HaKRkyKiGxJcUZSv5Uw6d4PgZuBvYCPVZoSoxqy2SxPv7iG\nsU11akGIiGxhcR77+QFhQr13AxuBfYAvJhlUXBs706zd0M2OW7dQU5Oq/AIREYktToKocfe7CBPr\nXevuzxFzkr+kvdAWRlBvNb6pwpkiItJXcRLEBjP7LHA4cLOZfQoYFDNgPfzUcgC2m6r7DyIiW1qc\nBHEyMAY4wd1XAVsD8xKNKqbHnw1PMO26ncZAiIhsaXEWDHoRuA6oNbODgFuAGUkHFkf9qFoAxo9t\nGOBIRESGnzhPMf2IMFhuKfklQbNAxdlck9bV3UNjfa1uUIuIJCDOzeajAHP3jUkH01cbu9I01tcO\ndBgiIsNSnHsQS3n9KOpBozud2dTNJCIiW1acFsRK4HEz+wfQkdvp7qclFlVMq9d1aYCciEhC4iSI\n26I/g0p3OvO6v0VEZMuqmCDc/Woz2x7YHfgTsI27L0s6sEo2dqUBmDSucYAjEREZniregzCz9wA3\nEdaFmAjca2bvTzqwStZt6Aagvk6LBImIJCFOF9MXgLnA3e7+mpntBfwVWFDphWaWAi4DZhHuX5zu\n7kujY1OA35KfSnxP4AvufkWcwLvSPQCMG10f53QREemjOF+/e9x909Qa7v4yELfj/zigwd3nAucC\nFxWU86q7H+ruh0XHFgFXxg18fYe6mEREkhSnBfGYmZ0N1JnZnsDHgYdjln8A0Q1ud7/PzPYtcd6l\nwDx3z5Y4/gadXaEFMbphUMwbKCIy7MRpQZwFTCdM9f0zYC0hScTRAqwp2E6b2eve08yOAR519yUx\nywTyCaJulO5BiIgkIc5TTOsJXUDnmtkkYGUfvumvBQqnWq1x997dU+8HLo5ZHq2tobie1KsATJ86\nbtO+kWakfu5iVBd5qos81cXmKZkgzKwV+DFhJbm7CBP2HQW8ambHuPvjMcpfSFjX+lozmw0sLnLO\nvu5+b9yA29rC7ZBVa8LMH12dXZv2jSStrc0j8nMXo7rIU13kqS7y+psoy7UgLgUeiP78F7A3Yarv\nnQiPvB4Zo/zrgSPNbGG0faqZzQPGuPt8M5vM67ugYuvqjrqYatXFJCKShHIJYjd3fy+Amb0duMbd\n1wIPmtnWcQqPuqLO7LX7yYLjywmJp886OqOb1I11/Xm5iIhUUO7rd+F9hsMIYx9yBnwCpO6ecCtj\nVO2gnEdQRGTIK9eCeDYaRT06+nMnQDSK+rHkQysvHSUIdTGJiCSjXII4C7gcmAK8z927zOwiwuJB\n76hGcOXkEsQoPeYqIpKIkgnC3Z/njYnga8DnijyqWnXpntADNkotCBGRRJS8uprZz8xs58J97r4q\nlxzMbHczuyrpAEtJ6x6EiEiiynUx/TdwsZlNA+4BXgDSwHbAodH2OYlHWEJuHQi1IEREklGui+lF\n4CQzm0EY7DaTMEnf08DJ7v50dUIsLt2TIZWC2hq1IEREkhBnqo2nCQPjBpWOrh4a60eRSilBiIgk\nYcj2z3R0pWmsrx3oMEREhq0hmSAy2Sxtqzs01beISIJiXWHNbAwwgzDZ3uhohtcBs7q9E4DlazsG\nMgwRkWEtzprUhwOPADcAU4FnzOyopAMrJ/cE027bTRjIMEREhrU4XUzfIKwMtzpabvRg4LuJRlVB\nLkFMbNZyoyIiSYmTIGrc/ZXcRsx1IBK1aaK+UXqCSUQkKXHuQbxgZkcDWTMbT5ij6blkwyovP4p6\nSN5jFxEZEuJcYc8ATga2IQyS2xP4SJJBVZLrYtJMriIiyYnTgpjl7vMKd5jZCcAfkgmpsk1TfWsm\nVxGRxJRbk/o9QANwoZl9pddrzmMAE0R3WjO5iogkrVwLogWYCzQTJufLSQNfSjKoSrp7wnKjWgtC\nRCQ55SbruxK40swOd/fbqxhTRemoBaF7ECIiyYlzD6LTzG4AxgIpoBbYzt23r/RCM0sBlwGzgA7g\ndHdfWnB8P+B70eYrwPvdvatSuWk95ioikrg4X8HnA/9LSCY/Ap4Cro9Z/nFAg7vPBc4FLup1/Arg\nQ+5+EHAbYa2JivQUk4hI8uJcYTe6+1XAncAqwiOuB8cs/wDChR93vw/YN3fAzHYBVgDnmNmdwER3\nfypOoRoHISKSvDhX2A4zmwg4MNvds8CYmOW3AGsKttNmlnvPycAc4AfAEcARZnZInEI3tSB0k1pE\nJDFx7kFcBPwOOAG438xOBhbFLH8t4SmonJrcmtaE1sMSd38SwMxuI7Qw7ixXYGtrM3XRNN+TJ42l\ntbW53OnD2kj+7L2pLvJUF3mqi80TZ0W535vZte6eNbN9gF2AJTHLX0hYrvRaM5tNmC48Zykw1sx2\njG5cH0i431FWW1s7a9vDNN/r13XQ1tYeM5ThpbW1ecR+9t5UF3mqizzVRV5/E2W5gXKtwDnASuD7\nhPEPGwljI24DpsQo/3rgSDNbGG2fambzgDHuPt/MPgz8xswA/uHuf4wTdK6LSfcgRESSU64F8Sug\nnXCvoN7MbgV+CYwGPhOn8Oh+xZm9dj9ZcPxOYP8+xAtoqg0RkWood4Wd4e4nErqI5gE3AwuAme7+\n62oEV0p+qg2NgxARSUq5FsRaAHdvj55iOtHd761OWOVtakGoi0lEJDHlrrDZgp9fHSzJAQpHUitB\niIgkpVwLotnMDiQkkTHRz5v6dNz97qSDK6VbA+VERBJXLkG8AFwY/fxiwc8QWheHJRVUJWlNtSEi\nkrhys7keWurYQOvuyVBbk6KmRjepRUSSMiS/gqfTWXUviYgkbEheZdM9GT3iKiKSsCGZILp7MnqC\nSUQkYRXnYjKzCcB3gBnAScB3gc+6+6qEYyupO53RDWoRkYTFucpeCdwPTCJMvfEyYUT1gEn3ZDTN\nhohIwuJcZXdw9yuAjLt3ufuXgDclHFdZ4R6EEoSISJLiXGXTZjaOaGS1me0MZMq/JFndeopJRCRx\ncRYMOp+wiM+2Zva/hFXgTksyqEp69BSTiEji4iSIvwAPEKblrgXOcPdXE42qjEw2Sxao1SA5EZFE\nxUkQzxEW/lng7v9MOJ6KMpkwh6BGUYuIJCtOgtgDOBH4HzObDvyWkCziLju6RW1KECklCBGRJMVZ\nk3oVYa3o+Wa2L3A58OU4r01CJqsWhIhINcQZKNdKGCD3XmAi8Gvg+ITjKikTPT+lFoSISLLitAIe\nBq4BPuPuixKOp6JcC0L5QUQkWXESxDbu3q9xD2aWAi4DZgEdwOnuvrTg+KeB04HXol1nuPtT5crs\n6EoD0FBX25+QREQkppIJwswedPe9CQPlCpcfTQFZd49zhT4OaHD3uWa2P3BRtC9nH+AD7v5Q3ICj\nBoQecxURSVi5BYP2jv5+w5BlM2uIWf4BwG1ROfdFN7kL7QOca2bTgFvc/VuVCszmupiUIEREElVx\nvgozu7fXdg1h4FwcLcCagu109Pqc3wAfAw4FDjCzd1QqMNeCUH4QEUlWuS6mvwGHRD8X3oNIAzfG\nLH8t0FywXdPrfsYl7r42eo9bgL2AW8sVOG78aACamuppbW0ud+qwN9I/fyHVRZ7qIk91sXnKdTEd\nBmBml7j7p/pZ/kLgaOBaM5sNLM4dMLMW4FEzmwlsBA4DflqpwBUr1wPQ2Zmmra29n2ENfa2tzSP6\n8xdSXeSpLvJUF3n9TZTlWhBHu/vNwINm9sHex939FzHKvx440swWRtunmtk8YIy7zzezcwkTAXYA\nt7v7bZUKzN2DUBeTiEiyyj3muh9wM1E3Uy9ZoGKCcPcscGav3U8WHP8V8KuKURa+cXQPIqWBECIi\niSrXxXR+9PepuX1Rt9A27v5YFWIrKquBciIiVRFnqo0PA28DvgA8BLSb2XXu/uWkgyums7sHgPpR\nGignIpKkOMuyfRz4HDAPuAF4M/CfSQZVTnc6PARVX6cV5UREkhTrKuvuK4F3EAazpYGmRKMqI90T\nupi05KiISLLiXGUfM7ObgR2Bv5rZNcD9yYZVWk9PaEGM0mNMIiKJipMgTgO+A+zv7l3ALwkT7A2I\ndLRgUK1aECIiiYpzla0nDHb7i5k9TBjQFncupi0unWtB1KoFISKSpDgJ4ofAaEJL4hSgDvhJkkGV\nk08QakGIiCQpznoQ+7j7rILts83s8aQCqkQ3qUVEqiPOVbbGzMbnNqKf08mFVJ66mEREqiNOC+Ii\n4H4zy83geizwzeRCKq+nRzepRUSqoeJV1t2vAo4HlgLPACe4+88SjqsktSBERKqj3GyuNcBZwC7A\nPe7+o6pFVcamBFGjFoSISJLKXWUvA04C1gPnmdlXqhNSebpJLSJSHeWusgcDB7v7FwljH06sTkjl\n5VoQtepiEhFJVLkE0RGt54C7ryCsATHgcgmiTi0IEZFElbvK9k4ImaJnVVn7hm4A6us13beISJLK\nPea6nZn9rNS2u5+WXFil5ab7bm6qG4i3FxEZMcoliHN6bd+VZCBx6TFXEZHqKLfk6NXVDCSu/E1q\n3YMQEUlSnJHU/WZmKcLjsrOADuB0d19a5LzLgRXufl6lMjd29VA/qoYaLUotIpKopL+GHwc0uPtc\n4FzCtB2vY2ZnAHvELTDdk6FulFoPIiJJi9WCMLMxwAxgMTDa3dfHLP8A4DYAd7/PzPbtVe4cYD/g\ncmBmnALT6YwGyYmIVEHFK62ZHQ48AtwATAWeMbOjYpbfAqwp2E5HU3hgZlOB84Gzgdj9RRs609TX\nKUGIiCQtTgviG4SWwB/d/WUzOxj4DfDnGK9dCzQXbNe4e248xUnAJOBWYBrQZGZPuPsvyhW4sbOH\nrSePpbW1udxpI4LqIE91kae6yFNdbJ44CaLG3V8xMwDc/fHczzEsJCxXeq2ZzSZ0UeXKuRS4FMDM\nTgGsUnIA6OnJkMlkaGtrjxvDsNTa2jzi6yBHdZGnushTXeT1N1HGSRAvmNnRQDZaLOgs4LmY5V8P\nHGlmC6PtU81sHjDG3ef3NdieTJYsUFujJ5hERJIWJ0GcAVwCbENYE+J24KNxCo/mcjqz1+4ni5wX\na8xFd7oH0BgIEZFqqJgg3P01YF4VYqloY0dY6XR0Q6LDN0REhBgJwsyWUWQmV3ffMZGIyujsDi0I\nPcUkIpK8OF/FDyn4uY6w/GhDItFUsLEztCAa6jSTq4hI0uJ0MT3ba9d3zewB4OvJhFRafqI+tSBE\nRJIWp4vpoILNFLA70JRYRGVkMqGnq0ZPMYmIJC5OF9MFBT9ngeXAKcmEU142uhOiefpERJIXJ0Fc\n4+4/TjySGHpyLQhlCBGRxMXpzD8r8ShiykZNiJQShIhI4uK0IJ43s78B9wEbczvd/cLEoiohk821\nIKr9ziIiI0+cBPHPgp8H9NKcjab5UxeTiEjySiYIMzvF3a929wtKnVNtuXsQKTUhREQSV+4exKeq\nFkVMnd3RQDmtKCcikrghdaXN5B5zVQtCRCRx5e5B7G5mS4vsTwHZgZiLKTcjlNKDiEjyyiWIJcA7\nqhVIHFn0mKuISLWUSxBdReZhGlAaSS0iUj3l7kEsLHNsQGwaKDfAcYiIjAQlE4S7n13NQOLItSDU\nhBARSd6QeopJ+UFEpHqGVoJQF5OISNUkurizmaWAy4BZQAdwursvLTh+IvAFIAP82t1/UK68/E1q\npQgRkaQl3YI4Dmhw97nAucBFuQNmVgN8AzgMmAt83Mwmli9OLQgRkWpJOkEcANwG4O73AfvmDrh7\nBtjV3dcBk6NYusoVlokm61OGEBFJXtIJogVYU7CdjloOQEgSZnY88DBwJ7C+fHG5FoQyhIhI0hK9\nBwGsBZoLtmuilsMm7n49cL2ZXQ18ELi6VGG5exAtLY20tjaXOm3EUB3kqS7yVBd5qovNk3SCWAgc\nDVxrZrOBxbkDZtYM3AQc5e5dhNZDpmgpkdxkfevWddDW1p5QyENDa2vziK+DHNVFnuoiT3WR199E\nmXSCuB440sxyo7JPNbN5wBh3n29mC4C7zawL+BewoFxh6Z6QP2o0m6uISOISTRDungXO7LX7yYLj\n84H5ccvrYI/nAAALzElEQVTr7ArrQTTVJ53XRERkSA2U60qHFkSdFgwSEUnckLrSbuiIVpSrqx3g\nSEREhr8hlSB6ooEQo2qHVNgiIkPSkLrSaj0IEZHqGWIJIlv5JBER2SKGVILIzfddoyaEiEjihlSC\nyGzqYxrYOERERoIhlSDyCwYpQ4iIJG1oJQg1IEREqmaIJYhoNldlCBGRxA2pBJGjLiYRkeQNqQSR\nyWhFORGRahlSCWITZQgRkcQNqQSRH0mtDCEikrQhlSBy4yCUHkREkjekEkSOEoSISPKG1Mo7c948\njTXtHUxsaRzoUEREhr0hlSBm7zGNGVPGDnQYIiIjwpDsYhIRkeQpQYiISFGJdjGZWQq4DJgFdACn\nu/vSguPzgE8B3cBid/94kvGIiEh8SbcgjgMa3H0ucC5wUe6AmTUCFwIHu/uBwHgzOzrheEREJKak\nE8QBwG0A7n4fsG/BsU5grrt3RtujCK0MEREZBJJOEC3AmoLttJnVALh71t3bAMzsE8AYd/9rwvGI\niEhMST/muhZoLtiucfdMbiO6R/EdYGfghBjlpVpbmyufNUKoLvJUF3mqizzVxeZJOkEsBI4GrjWz\n2cDiXsevADa6+3EJxyEiIn2Uyi3Ck4SCp5jeEu06FdgHGAMsAu4H/h4dywKXuPsNiQUkIiKxJZog\nRERk6NJAORERKUoJQkREilKCEBGRogblbK4xpug4BvhvwhQdV7n7/AEJtAo0XUlepbooOO9yYIW7\nn1flEKsmxu/FfsD3os1XgPe7e1fVA01YjHo4GTgHSBOuFT8ZkECryMz2B77l7of22t/n6+ZgbUGU\nm6JjVLR9BHAI8FEzax2IIKtE05XklayLHDM7A9ij2oENgEp1cQXwIXc/iDCbwXZVjq9aKtXDd4HD\nCLM6fNbMxlU5vqoys88DVwINvfb367o5WBNEuSk6dgWecve17t4N3AMcVP0Qq0bTleSVqwvMbA6w\nH3B59UOrupJ1YWa7ACuAc8zsTmCiuz81EEFWQdnfCeARYALQFG0P98c2lwDHF9nfr+vmYE0QJafo\nKHKsHRjO3wo0XUleybows6nA+cDZjIxVacv9H5kMzAF+QPjGeISZHVLd8KqmXD0APEYYc7UYuNnd\n11YzuGpz9+sJ3Wm99eu6OVgTRLkpOtYSPmxOM7C6WoENgIrTlZjZd4HDiTddyVBWri5OAiYBtwJf\nBN5nZh+scnzVVK4uVgBL3P1Jd08TvmH3/mY9XJSsBzN7M/BOQvfa9sAUMzux6hEODv26bg7WBLEQ\neAdAkSk6/g3sZGbjzaye0Ey6t/ohVk25uoDQ19zg7scVdDUNVyXrwt0vdff93P0w4FvAr939FwMT\nZlWU+71YCow1sx2j7QMJ36SHo3L1sAbYAHS6exZ4jdDdNBL0bkX367o5KEdSl5uiw93nm9k7Cd0J\nKeCnw/nJBE1Xklfp96LgvFMAGyFPMZX6P3II8O3o2D/c/TPVjzJ5MerhDOA0wv26p4GPRK2qYcvM\ntgN+4+5zo6cc+33dHJQJQkREBt5g7WISEZEBpgQhIiJFKUGIiEhRShAiIlKUEoSIiBSlBCEiIkUN\nytlcpXqiZ6afJD+QKkUYT3GMu79Y4jXnA1l3v3Az3vcUwuRhz0bv2QjcBXy8cKR4zLIuAO5395vN\n7G/RYDnM7EF337u/MUZl3AG8iTA1QYowGvVp4OTcNCclXvcRYK27/64P7zUd+Jq7n1aw70IgXaqu\no9HCFxNGkdcSBj992t03xH3fGHHdDJxOGGj2R2Br4Cpgprt/tMRr9gHOcPePVqoLMxsD/AJ4dzSg\nTQYJJQgBeHFzL6T9dEPuYhgNeLoLOAu4tC+FuPv5BZuHFOzfUp/pNHfPDUbEzK4jTCF9bpnXzAXu\n6OP7XAx8KXqPFkICfS/wnTKv+R1h1tb/i153GWGG38/18b1Lcvejo7K3BXZ39zfFeM0iIJc8ytaF\nu683s78AHwN+vPkRy5aiBCElmdnuhIv1GGAr4Hvu/sOC46OAnwG7R7t+HI3Y3Iowo+qbgAxwnrvf\nXu693D1rZv8AdonKPpVwEc4QRoyfDXT1er/L3P2nZnYVcCewd/Tae919jpllCL/jzwN7unubmU0A\nHgW2BY4ELojOWUYYZbuqSHibumLNrJkwGd4/o+2TojgbCTOGnk6YavlY4FAze5kwo2jZ+jCzGcA0\nd38y2vUuQsvue5Q3hfDvk/NVwrxDRPWSAd5MaPl83d0XRN/YfxTVYy3wbXf/nZk1RPsPINT119z9\n92a2DDgYuAmYbGb/B3we+Kq7H2pmewI/iT7/SuD9wE5RLF8vqIvVwE+BHdx9XdR6vcXd9yAkun+i\nBDGo6B6EAEw3swfN7KHo789G+08nXCT2J8yp/41er5tLmEp6H8LFdm60/xLCUP79CBe6y6OLUklm\nNgl4O3CPme0BnAcc6O6zCPPpfLXI+72toIisu38KwN3nFOzLANcQJvMDOBG4njAnzzeBo6Ly/kzp\nb+pXRnXzEqEL58/A96NWz0eBd7r7XoSpLT4fXfxvBL7i7n+JWR9HE6ZgJvoMv3T37xAu8OV8BrjJ\nzDxaKGnfXGsiMh2YTZjM8f9FyfvLwANRPAcDXzaz7YHcjMAzCfX7FTOrKyjrWOAld39rtJ3rDloA\nXBD9W/0W+GTueK+6uBG4GXh3dPyDwNXR510FtEddZjJIqAUhULqL6bPAf5rZFwlz3fS+qD0K7GJm\ntxFmUf1CtP8IwMzsa9F2LTAD+Fev17/LzB4kfFFJAddF32TPAm5099xsk1cQWg7fLPF+lSwAvk+Y\ns2ceoRtnf0Ir4o7oQl9DmAW1mA+7+9+j9SauBW7NzedjZicAx5iZEbq3is3zE6c+dgaeiPl5NnH3\nX0RdXkdEf64ys1+5+znRKVdFSfJFM7uHMHHfEUCTmX04OqeJ0Jo4mGgtDXd/ldDyIHy04qLEPtXd\n/xi97vJo/8ElXnIVYT6gnwPvAwpXPXuOUA+9J6SUAaIEIeX8nnDRvInwzfA9hQfdfWX0bf8IwrTK\nD0XdUjXAYbkLvJlNIyx72dumexC99G7ZpoBR7r6qyPvtVulDuPsiM5toZvsC0939n2Z2LPB3dz8u\nirGe108b3fv9cfd7zexS4Jdm9hbChfV+wg3WuwgX/LNKfJ5K9ZGheHLZJHrdrYRv7i8Bnwbe6+5f\nB24AbjCzS4CHCN1e9CqzlrDcZA1hCdKHo3K3InQNfbjg3Fy313PlYorK2zRzaNRNtXWpk939bjOb\nbmbHA0vdvbAeuqncYpIqUheTQOkFdg4ndA3cRHTzN/q2TfTzMcACd7+VsC52O6Gf/W9EF8roAv4v\nYHQf4rkTONbMxkfbHyF80y/2ftv0em3hgjGFn+vXhG/Hv4227wPmmNnO0fb5hOUpK7ko+ixnEu6X\n9Lj7Nwg3Yd9OuAhDuDDnvoDFqY+nqbAsqLu/7O57ufve0Y3jNuCT9vrFgPYgJIic/4redzvgrYSZ\nf+8APh7tnxbFsw1wd8H5WxH+HV63dCW9fleiBXieM7PDo10fJNzXKZQGCruqfkFYzOiqXuftQFgR\nTQYJJQiB0sswfhVYaGYPEPqklxH+E+fcCmw0s8cINxivc/fHCH3Qs83sEeA3hEdC18cNxt0XE7qT\n7jazxwkrX32Z8IjlhiLvVxj/jcAj0TfZwv0LCAvbL4je41XCNNDXRHHuSehS6+11dePuXVEsXyFc\nzB4xMyfcSG8nf5H/K3Be1AX1iRj1cTOv726pyN3XEFpSXzWzJVFdnULoRssZHf373UT+JvwFhC6m\nxVGcn3P3ZYQuuA1RnH8Gznb3db3qoNjvygeiGB4k3Ov5fK/jfwXOjeoCwg3pJkKrBwALa0W3uPuj\nfakDSZam+xYZJMzsWuD8KOltifKuAu4YTAsnRS3QM4Fd3P3TBfs/CXS7u55iGkR0D0Jk8DiH8O3+\n1C1U3mD89vcHQnfWf+R2RE90HQ4cP1BBSXFqQYiISFG6ByEiIkUpQYiISFFKECIiUpQShIiIFKUE\nISIiRSlBiIhIUf8fyxroZmNK+KYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x132168c10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgYAAAHBCAYAAAASOnMxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIEmSKgYGkiSpYmAgSZIqBgaSJKny/wGL88QTLtzJgwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13bffebd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature Importance: \n",
      "\n",
      "0.7105 \t\tnum_sessions\n",
      "0.0583 \t\ttotal_artists_followed\n",
      "0.1803 \t\ttotal_artworks_favorited\n",
      "0.0175 \t\ttotal_artworks_shared\n",
      "0.0094 \t\tgender_male\n",
      "0.0239 \t\tgender_unknown\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdgAAAEZCAYAAAA0bTSIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13bcf0c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = fit_random_forest(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 517,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "a = np.array([[ 0.18966242,  0.37135351,  0.10676733,  0.56563498,  0.        ,  1.        ], \n",
    "             [ 0.1966242,  0.47135351,  0.19676733,  0.06563498,  1.        ,  0.        ],\n",
    "             [-0.15949266, -0.27305311, -0.23691142, -0.11191526,  0.        , 0.        ]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 780,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.0265,  0.9735],\n",
       "       [ 0.087 ,  0.913 ],\n",
       "       [ 1.    ,  0.    ]])"
      ]
     },
     "execution_count": 780,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict_proba(a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 779,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.97350000000000003, 0.91300000000000026, 0.0]"
      ]
     },
     "execution_count": 779,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[i[1] for i in probs]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Price distribution (ticket):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 785,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.3075905 ,  1.76841926,  0.03038993,  0.56564144,  1.        ,  0.        ])"
      ]
     },
     "execution_count": 785,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 763,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.81754167,  0.18245833])"
      ]
     },
     "execution_count": 763,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "probs[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 759,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "numpy.ndarray"
      ]
     },
     "execution_count": 759,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 762,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.29192688492063557"
      ]
     },
     "execution_count": 762,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Predicting on entire dataset:\n",
    "probs = model.predict_proba(X)\n",
    "probss = pd.Series([i[1] for i in probs])\n",
    "# probs.plot.kde(normalized=True)\n",
    "probss.quantile(q=0.93)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 791,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "p = model.predict_proba(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 828,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "matplotlib.rc('xtick', labelsize=13) \n",
    "matplotlib.rc('ytick', labelsize=13)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 860,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAakAAAFHCAYAAAAbYrOpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIjklpKUiIjk1v8HUQZtvxA4MdwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x139132c90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist([i[1] for i in p])\n",
    "plt.title(\"\\nPredictions distribution on original dataset\\n\", fontsize='small')\n",
    "plt.xlabel(\"Probability of an User making a purchase\", fontsize=13)\n",
    "plt.ylabel(\"Number of Users\", fontsize=13)\n",
    "plt.savefig('images/probs_of_purchase2.png', bbox_inches='tight', pad_inches=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 859,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaEAAAFYCAYAAAD6NUMTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x147a25190>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Zoomed in for better viz:\n",
    "plt.hist([i[1] for i in p])\n",
    "plt.ylim(0,3000)\n",
    "plt.title(\"\\nPredictions distribution on original dataset\\n[close-up]\\n\", fontsize='small')\n",
    "plt.xlabel(\"Probability of an User making a purchase\", fontsize=13)\n",
    "plt.ylabel(\"Number of Users\", fontsize=13)\n",
    "plt.savefig('images/probs_of_purchase_zoom2.png', bbox_inches='tight', pad_inches=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 790,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x140208e10>"
      ]
     },
     "execution_count": 790,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAECCAYAAAAVYxsVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13c02c150>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probss.hist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 789,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Total registered users: 57816\n",
      "\n",
      "Total registered users with 40-95% purchase probability: 2387, or 4.1% of total users. \n",
      "There will be more costs to convert users, but the potential lift considering full conversion would be of $637329.0 - or 60.7% of revenue.\n",
      "\n",
      "Total registered users with 60-95% purchase probability: 1267, or 2.2% of total users. \n",
      "There will be more costs to convert users, but the potential lift considering full conversion would be of $338289.0 - or 32.2% of revenue.\n",
      "\n",
      "Total registered users with 80-95% purchase probability: 532, or 0.9% of total users. \n",
      "There will be more costs to convert users, but the potential lift considering full conversion would be of $142044.0 - or 13.5% of revenue.\n"
     ]
    }
   ],
   "source": [
    "print \"\\nTotal registered users:\", probss.count()\n",
    "\n",
    "def recommender(threshold, ticket=267.):\n",
    "    _1 = probss[(probss>threshold) & (probss<0.95)].count()\n",
    "    _2 = round(probss[(probss>threshold) & (probss<0.95)].count()/float(probss.count())*100, 1)\n",
    "    _3 = ticket * _1\n",
    "    _4 = round(_3 / 1050000. *100, 1)\n",
    "    \n",
    "    print \"\\nTotal registered users with {}-95% purchase probability: {}, or {}% of total users. \\nThere will be more costs \\\n",
    "to convert users, but the potential lift considering full conversion would be of ${} - \\\n",
    "or {}% of revenue.\".format(int(threshold*100),_1, _2, _3, _4)\n",
    "\n",
    "for threshold in [0.4, 0.6, 0.8]:\n",
    "    recommender(threshold)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\n",
    "'''\n",
    "Recommendations: \n",
    "Focus on users with between 30% and 95% of purchase probability (representing 5-6% of users)\n",
    "\n",
    "Average ticket of $515; but distribution tell other story:\n",
    "Average ticket for the biggest chunk (bellow $1,500) is $314\n",
    "\n",
    "Strategies/scenarios\n",
    "1) Agressive:\n",
    " - Target users with probability of purchase between 30% and 95%\n",
    " - \n",
    "\n",
    "With a $120 ticket and about 4850 targeted users, potential lift = $378,000 (going from 30% prob until 95% prob)\n",
    "With the same ticket and about 1450 targeted users, potential lift = $174,000 (going from 50% prob until 95% prob - \n",
    "representing 16.5% increase in revenue)\n",
    "Super conservative (less expenses to convert users): focus on 70-95% probability of purchase\n",
    "(500+300+300+350+250) * 120\n",
    "\n",
    "'''\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Trying to fit profit curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 533,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def profit_curve(cost_benefit_matrix, probabilities, y_true):\n",
    "    thresholds = sorted(probabilities)\n",
    "    thresholds.append(1.0)\n",
    "    profits = []\n",
    "    for threshold in thresholds:\n",
    "        y_predict = probabilities >= threshold\n",
    "        confusion_mat = standard_confusion_matrix(y_true, y_predict)\n",
    "        profit = np.sum(confusion_mat * cost_benefit_matrix) / float(len(y_true))\n",
    "        profits.append(profit)\n",
    "    return thresholds, profits\n",
    "\n",
    "\n",
    "def run_profit_curve(model, costbenefit, X_train, X_test, y_train, y_test):\n",
    "    model.fit(X_train, y_train)\n",
    "    probabilities = model.predict_proba(X_test)[:, 1]\n",
    "    thresholds, profits = profit_curve(costbenefit, probabilities, y_test)\n",
    "    return thresholds, profits\n",
    "\n",
    "\n",
    "def plot_profit_models(models, costbenefit, X_train, X_test, y_train, y_test):\n",
    "    percentages = np.linspace(0, 100, len(y_test) + 1)\n",
    "    for model in models:\n",
    "        thresholds, profits = run_profit_curve(model,\n",
    "                                               costbenefit,\n",
    "                                               X_train, X_test,\n",
    "                                               y_train, y_test)\n",
    "        plt.plot(percentages, profits, label=model.__class__.__name__)\n",
    "    plt.title(\"Profit Curves\")\n",
    "    plt.xlabel(\"Percentage of test instances (decreasing by score)\")\n",
    "    plt.ylabel(\"Profit\")\n",
    "    plt.legend(loc='upper left')\n",
    "    plt.savefig('profit_curve.png')\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "def find_best_threshold(models, costbenefit, X_train, X_test, y_train, y_test):\n",
    "    max_model = None\n",
    "    max_threshold = None\n",
    "    max_profit = None\n",
    "    for model in models:\n",
    "        thresholds, profits = run_profit_curve(model, costbenefit,\n",
    "                                               X_train, X_test,\n",
    "                                               y_train, y_test)\n",
    "        max_index = np.argmax(profits)\n",
    "        if not max_model or profits[max_index] > max_profit:\n",
    "            max_model = model.__class__.__name__\n",
    "            max_threshold = thresholds[max_index]\n",
    "            max_profit = profits[max_index]\n",
    "    return max_model, max_threshold, max_profit\n",
    "\n",
    "\n",
    "def main(X, y):\n",
    "    sm = SMOTE(kind='regular', ratio=0.5)\n",
    "    X_resampled, y_resampled = sm.fit_transform(X, y)\n",
    "    \n",
    "    # Splitting train and test data\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3)\n",
    "    \n",
    "    costbenefit = np.array([[0, 120], [-20, 0]])\n",
    "    models = [model]\n",
    "    # plot_profit_models(models, costbenefit,\n",
    "    #                    X_train, X_test, y_train, y_test)\n",
    "    plot_profit_models(models, costbenefit, X_train, X_test, y_train, y_test)\n",
    "    print find_best_threshold(models, costbenefit,\n",
    "                              X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 531,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEZCAYAAACEkhK6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x131176fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('RandomForestClassifier', 0.0, 85.611603131818768)\n"
     ]
    }
   ],
   "source": [
    "main()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 532,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "\n",
    "# Dropping date columns due to high relationship and used to create the label:\n",
    "df.drop('diff_last_fav_artwork_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_last_artist_folw_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_created_to_last', axis=1, inplace=True)\n",
    "df.drop('days_from_last_session', axis=1, inplace=True)\n",
    "\n",
    "merged_df = preprocess_purchases_and_join_with(df)\n",
    "\n",
    "# Scaling features\n",
    "merged_df.num_sessions = scale(merged_df.num_sessions)\n",
    "merged_df.total_artists_followed = scale(merged_df.total_artists_followed)\n",
    "merged_df.total_artworks_favorited = scale(merged_df.total_artworks_favorited)\n",
    "merged_df.total_artworks_shared = scale(merged_df.total_artworks_shared)\n",
    "\n",
    "# Defining y label and X matrix\n",
    "y = merged_df.pop('purchased').values\n",
    "X = merged_df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "main(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model 6: Random Forests with Label = activity in last 30 days (ignoring 0 days of activity -> 17k records dropped)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "\n",
    "def negative_to_zero(x):\n",
    "    if x == -1:\n",
    "        return 0\n",
    "    else:\n",
    "        return x\n",
    "\n",
    "df.diff_created_to_last = df.diff_created_to_last.apply(lambda x: negative_to_zero(x))\n",
    "\n",
    "df = df[df.diff_created_to_last > 0]\n",
    "\n",
    "create_label(df, name='engagement30', original_col_name='diff_created_to_last', smaller_than=30)\n",
    "\n",
    "# Defining y label and X matrix\n",
    "y = df.pop('engagement30').values\n",
    "X = df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({0: 27433, 1: 13224})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 492 new samples ...\n",
      "done!\n",
      "Random Forest Scores:\n",
      "\n",
      "accuracy: 0.675820170109\n",
      "precision: 0.514166418133\n",
      "recall: 0.420795704174\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAECCAYAAAAb5qc/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXH8WAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1157477d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature Importance: \n",
      "\n",
      "0.466 \t\tnum_sessions\n",
      "0.1449 \t\ttotal_artists_followed\n",
      "0.2983 \t\ttotal_artworks_favorited\n",
      "0.0457 \t\ttotal_artworks_shared\n",
      "0.0187 \t\tgender_male\n",
      "0.0264 \t\tgender_unknown\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfEAAAGkCAYAAAAhcRw5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKENckqREGeKSJCXKEJckKVH/B7UUcR7PbbAQAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1199c89d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fit_random_forest(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Refining models by scaling and grid searching"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 887,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_importance(clf, X, max_features=10):\n",
    "    '''Plot feature importance'''\n",
    "    feature_importance = clf.feature_importances_\n",
    "    # make importances relative to max importance\n",
    "    feature_importance = 100.0 * (feature_importance / feature_importance.max())\n",
    "    sorted_idx = np.argsort(feature_importance)\n",
    "    pos = np.arange(sorted_idx.shape[0]) + .5\n",
    "    \n",
    "    # Show only top features\n",
    "    pos = pos[-max_features:]\n",
    "    feature_importance = (feature_importance[sorted_idx])[-max_features:]\n",
    "    feature_names = (X.columns[sorted_idx])[-max_features:]\n",
    "    \n",
    "    plt.barh(pos, feature_importance, align='center')\n",
    "    plt.yticks(pos, feature_names)\n",
    "    plt.xlabel('Relative Importance', fontsize=13)\n",
    "    plt.title('\\nVariable Importance\\n', fontsize='small')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "drop_cols = ['last_sign_in_at', \n",
    "             'created_at', \n",
    "             'last_session', \n",
    "             'last_favorited_artwork_date', \n",
    "             'last_followed_artist_date',\n",
    "             'total_follows',\n",
    "             'total_favorites',\n",
    "             'city']\n",
    "df = preprocess(drop_cols)\n",
    "\n",
    "# Dropping date columns due to high relationship and used to create the label:\n",
    "df.drop('diff_last_fav_artwork_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_last_artist_folw_to_today', axis=1, inplace=True)\n",
    "df.drop('diff_created_to_last', axis=1, inplace=True)\n",
    "df.drop('days_from_last_session', axis=1, inplace=True)\n",
    "\n",
    "merged_df = preprocess_purchases_and_join_with(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Defining y label and X matrix\n",
    "y = merged_df.pop('purchased').values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Scaling features\n",
    "merged_df.num_sessions = scale(merged_df.num_sessions)\n",
    "merged_df.total_artists_followed = scale(merged_df.total_artists_followed)\n",
    "merged_df.total_artworks_favorited = scale(merged_df.total_artworks_favorited)\n",
    "merged_df.total_artworks_shared = scale(merged_df.total_artworks_shared)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X = merged_df.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Grid Searching Random Forests:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Determining classes statistics... 2 classes detected: Counter({0: 55649, 1: 2168})\n",
      "Finding the 5 nearest neighbours...\n",
      "done!\n",
      "Creating synthetic samples...Generated 20091 new samples ...\n",
      "done!\n",
      "Fitting 3 folds for each of 324 candidates, totalling 972 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:   53.0s\n",
      "[Parallel(n_jobs=-1)]: Done 192 tasks      | elapsed:  4.5min\n",
      "[Parallel(n_jobs=-1)]: Done 442 tasks      | elapsed: 12.6min\n",
      "[Parallel(n_jobs=-1)]: Done 792 tasks      | elapsed: 27.8min\n",
      "[Parallel(n_jobs=-1)]: Done 972 out of 972 | elapsed: 38.5min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'oob_score': True, 'n_jobs': -1, 'bootstrap': True, 'min_samples_leaf': 2, 'n_estimators': 400, 'min_samples_split': 1, 'random_state': 67, 'max_features': None, 'max_depth': None}\n",
      "OOB: 0.954121206565\n"
     ]
    },
    {
     "ename": "AttributeError",
     "evalue": "'numpy.ndarray' object has no attribute 'columns'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-203-c406e79e7118>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     31\u001b[0m \u001b[0;32mprint\u001b[0m \u001b[0;34m\"OOB:\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbest_model\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moob_score_\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     32\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 33\u001b[0;31m \u001b[0mplot_importance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbest_model\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX_train\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmax_features\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m16\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-182-84aa9ca35097>\u001b[0m in \u001b[0;36mplot_importance\u001b[0;34m(clf, X, max_features)\u001b[0m\n\u001b[1;32m     14\u001b[0m     \u001b[0mpos\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpos\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mmax_features\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     15\u001b[0m     \u001b[0mfeature_importance\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mfeature_importance\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msorted_idx\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mmax_features\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 16\u001b[0;31m     \u001b[0mfeature_names\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcolumns\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msorted_idx\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mmax_features\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     17\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     18\u001b[0m     \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbarh\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpos\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeature_importance\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malign\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'center'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'numpy.ndarray' object has no attribute 'columns'"
     ]
    }
   ],
   "source": [
    "rf_grid = {\n",
    "    'max_depth': [4, 8, None],\n",
    "    'max_features': ['sqrt', 'log2', None],\n",
    "    'min_samples_split': [1, 2, 4],\n",
    "    'min_samples_leaf': [1, 2, 4],\n",
    "    'bootstrap': [True], # Mandatory with oob_score=True\n",
    "    'n_estimators': [50, 100, 200, 400],\n",
    "    'random_state': [67],\n",
    "    'oob_score': [True],\n",
    "    'n_jobs': [-1]\n",
    "    }\n",
    "\n",
    "rf_grid_cv = GridSearchCV(RandomForestClassifier(),\n",
    "                             rf_grid,\n",
    "                             n_jobs=-1,\n",
    "                             verbose=True,\n",
    "                             scoring='roc_auc')\n",
    "\n",
    "sm = SMOTE(kind='regular', ratio=0.4)\n",
    "X_resampled, y_resampled = sm.fit_transform(X, y)\n",
    "\n",
    "# Splitting train and test data\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.3)\n",
    "\n",
    "rf_grid_cv.fit(X_train, y_train)\n",
    "\n",
    "print rf_grid_cv.best_params_\n",
    "\n",
    "best_model = rf_grid_cv.best_estimator_\n",
    "\n",
    "print \"OOB:\", best_model.oob_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdgAAAEZCAYAAAA0bTSIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115a37e50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_importance(best_model, merged_df, max_features=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "42"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Trying to understand puchase cycles:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 869,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "After the first purchase, how many days until purchase #2? 108.612149533 days\n"
     ]
    }
   ],
   "source": [
    "# Once a user makes a purchase, how long does it take for him to buy again?\n",
    "\n",
    "pc = pd.read_csv('purchase_cycles.csv')\n",
    "pc.when = pd.to_datetime(pc['when'])\n",
    "\n",
    "# Dropping Vango's team ids\n",
    "vango_ids = [38175, 1, 1326, 587, 736, 45651, 67966, 48516, 84261, 30975, 4260]\n",
    "for _id in vango_ids:\n",
    "    pc = pc[pc.id != _id]\n",
    "\n",
    "pc = pc.sort_values('when', axis=0)\n",
    "pc['diffs'] = pc.groupby('id').when.diff()\n",
    "pc['diffs'] = pc.diffs.apply(lambda x: extract_days(x))\n",
    "pc = pc[pc.diffs != 0]\n",
    "\n",
    "print \"After the first purchase, how many days until purchase #2?\", pc.diffs.mean(), \"days\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 865,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Unnamed: 0      955.5\n",
       "id            47230.0\n",
       "pieces            1.0\n",
       "spent           250.0\n",
       "diffs            48.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 865,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pc.median()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 569,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "103.07000000000001"
      ]
     },
     "execution_count": 569,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Average ticket\n",
    "pc.total_spent.mean()\n",
    "515.35*0.20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 687,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "266.65559932470455"
      ]
     },
     "execution_count": 687,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pc[pc.total_spent<1000].total_spent.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 871,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'Unnamed: 0', u'id', u'pieces', u'spent', u'when', u'diffs'], dtype='object')"
      ]
     },
     "execution_count": 871,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pc.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 875,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ8AAAFHCAYAAAB+j0HWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x140cf2fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pc.spent.hist(bins=20)\n",
    "plt.xlim(0,5000)\n",
    "plt.title(\"\\nPurchases ticket/price\\n\", fontsize='small')\n",
    "plt.xlabel(\"Total spent in purchases\", fontsize=13)\n",
    "plt.ylabel(\"Number of Users\", fontsize=13)\n",
    "plt.savefig('images/purchases_ticket2.png', bbox_inches='tight', pad_inches=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Compute difference between signup and purchase dates.\n",
    "#Extract the number days (don't carea bout precise millisecond)\n",
    "#groupby sum on number of days for num of purchases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 877,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdcAAAEXCAYAAAD/bA51AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13bdebad0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Loading csv extracted from database into a Pandas DataFrame\n",
    "df = pd.read_csv('dataset.csv', usecols=range(1,17))\n",
    "\n",
    "# Treating date fields and converting them to datetime timestamp\n",
    "df['acc_created_at'] = pd.to_datetime(df['created_at'])\n",
    "df.drop('created_at', axis=1, inplace=True)\n",
    "df.last_session = pd.to_datetime(df['last_session'])\n",
    "df.last_favorited_artwork_date = pd.to_datetime(df['last_favorited_artwork_date'])\n",
    "df.last_followed_artist_date = pd.to_datetime(df['last_followed_artist_date'])\n",
    "\n",
    "# Creating function to extract only the days (int) from TimeDelta objects\n",
    "def extract_days(x):\n",
    "    try:\n",
    "        return x.days\n",
    "    except:\n",
    "        return 0\n",
    "\n",
    "def negative_to_zero(x):\n",
    "    if x < 0:\n",
    "        return 0\n",
    "    else:\n",
    "        return x    \n",
    "\n",
    "df = df.set_index('id')\n",
    "\n",
    "# Loading data\n",
    "purch = pd.read_csv('purchases.csv', usecols=range(1,5))\n",
    "\n",
    "# Dropping Vango's team ids\n",
    "vango_ids = [38175, 1, 1326, 587, 736, 45651, 67966, 48516, 84261, 30975, 4260]\n",
    "for _id in vango_ids:\n",
    "    purch = purch[purch.user_id != _id]\n",
    "\n",
    "purch = purch.set_index('user_id')\n",
    "\n",
    "# Dropping purchases from anonymous non-registered users\n",
    "purch.dropna(inplace=True)\n",
    "\n",
    "pc = df.join(purch)\n",
    "\n",
    "pc.created_at = pd.to_datetime(pc['created_at'])\n",
    "\n",
    "# New feature: Extracting the difference between last session and today\n",
    "pc['diff_acc_to_purchase'] = pc.created_at - pc.acc_created_at\n",
    "pc['diff_acc_to_purchase'] = pc.diff_acc_to_purchase.apply(lambda x: extract_days(x))\n",
    "pc['diff_acc_to_purchase'] = pc.diff_acc_to_purchase.apply(lambda x: negative_to_zero(x))\n",
    "\n",
    "pc = pc[pc.diff_acc_to_purchase != 0]\n",
    "\n",
    "\n",
    "pc = pc.sort_values('diff_acc_to_purchase')\n",
    "\n",
    "with plt.style.context('fivethirtyeight'):\n",
    "    plt.plot(pc.diff_acc_to_purchase, np.cumsum(pc.total_spent))\n",
    "    plt.xlim(-50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 559,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "112.32029702970297"
      ]
     },
     "execution_count": 559,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pc.diff_acc_to_purchase.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 879,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAFYCAYAAAC77fzpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n85w83Ls3jDsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1483e1bd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(pc.diff_acc_to_purchase, np.cumsum(pc.total_spent))\n",
    "plt.title(\"\\nCumulative Revenue centering all\\n users on their Registration Day\\n\", fontsize='small')\n",
    "plt.yticks([])\n",
    "plt.xlabel(\"Days since each User registered\", fontsize=13)\n",
    "plt.ylabel(\"Cumulative Revenue\", fontsize=13)\n",
    "plt.savefig('images/cumulative_revenue_from_registrations2.png', bbox_inches='tight', pad_inches=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x133877fd0>]"
      ]
     },
     "execution_count": 298,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAAECCAYAAADdD/HDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3X2clOV97/HPzOzO7rLMAsKACqig8UdMlaipaIqoUY5P\nTT2mp41aU2NSrR5iWnOS04QcPZaWqEn1EBNDWxNfMRrTV2Jqjo1Hxdaoi5oIRqM05IfErKI8LSyw\nu+zj7Nznj/teHFZgh+HenRn4vl/6mplrrrn2N/cu85376boTQRAgIiISl2S5CxARkYOLgkVERGKl\nYBERkVgpWEREJFYKFhERiZWCRUREYlVTTCczmwPc5u7nFLRdAXzG3T8cPb4GuBboBxa7+6NmVg88\nAEwG2oGr3H2rmZ0OLIn6Punui6IxbgYujtpvdPcVZjYReBCoB9YDV7t7TwzvXURERsCwayxm9gXg\nHqCuoO1k4FMFj6cANwBnABcAt5pZLXA98Kq7zwPuB26KXrIUuMzdzwTmmNnsaMx57j4HuBy4O+p7\nM/B9dz8LeAW47gDer4iIjLBiNoWtBS4dfBCtQfw98FcFfU4Dlrt7zt3bgdeB2cBc4PGoz2PAuWaW\nAdLu3hK1PwHMj/ouA3D3dUDKzCbtaYz9fI8iIjKKhg0Wd38YyAGYWRL4NvA5YGdBtyZgR8HjTmAc\nkClo7yhoay/o27GHvntrH2wTEZEKVdQ+lgKnAMcRbspqAN5vZncCPyMMl0EZYBthgGQK2rYThsOe\n+vYV9CXqUzhGb8EYIiJSofYnWBLuvhI4EcDMjgZ+4O6fi/ax/L2ZpQkDZxawCngeuAhYGd02u3uH\nmfWa2QygBTgfuAUYAG43szuA6dHPazOz56LXfg+4EGguptggCIJEIrEfb09E5JAXy4fm/gTLXmer\ndPdNZnYXsJywsIXu3mdmS4H7zKyZcI3jiugl1xEe6ZUElrn7CoCo3wvRGAuivoujMa4BthSMsU+J\nRILW1o79eHujL5vNVHyNoDrjpjrjpTrjk81mhu9UhMRBPLtxUA2/xEqvEVRn3FRnvFRnfLLZTCxr\nLDpBUkREYqVgERGRWClYREQkVgoWERGJlYJFRERipWAREZFYKVhERCRWChYREYmVgkVERGKlYBER\nkVgpWEREJFYKFhERiZWCRUREYqVgERGRWClYREQkVgoWERGJlYJFRERipWAREZFYKVhERCRWChYR\nEYlVTbkLEBHJ5wO6+3IEAeSDILzNBwSD94Pd7/f2D9C6vYcgCBi7bgc72ruH9GXX4yAIyO/pMcM8\nHwTvGetAxk6na+jp6X/vOPtaMME+n933a4d5eTIBF8w5mg/MOGyYUfafgkVERkw+CNje0cvmbd1s\n6+ilNzdAe2cf2zp7GcgH9PTm2NjWzca2LnID+XKXe0hJACfOnKhgEZHKkBvI05/L05fL098/QG8u\nT3tnL2vXt/O79e1saOtiYCDP9s6+YQMjXZtkaraRCWPrSCYTJBOQSCRIJCCZTJDg3bZkMrxNJRNM\nmTCGmlSCTFMDOzt73n1NdLv74wQJojEKnivm8W5jUvA4uYfXsvexstkMbVs7dx876r8vwzw9rH29\nPpUcmb0hChYR2au3WztZ9UYbff0D9OXy/PvKdfQP5IfbQsPYhlpqUgmmZRuZPKGB7PgGJjbVU5dO\nUZtKkkomOGJSIw3pFJnGNMkD+PTMZjO0tnaU/PrRkhmTpmdnbbnLGBVFBYuZzQFuc/dzzOyDwF1A\nDugF/tzdW83sGuBaoB9Y7O6Pmlk98AAwGWgHrnL3rWZ2OrAk6vukuy+Kfs7NwMVR+43uvsLMJgIP\nAvXAeuBqd++JawGIHKq27Ohme2ffrjWO/lyevv4Btrb3hJuuOvtY3dK2x9e+/+gJ1NWmqK1Jkq5J\nUluboiGd4tip4zj8sDEcOalxlN+NVJJhg8XMvgB8AuiMmpYAC9z9NTO7FvgbM/sacANwCjAGWG5m\ny4DrgVfdfZGZfRy4CfhrYClwqbu3mNmjZjab8Ai1ee4+x8ymAz8GTgNuBr7v7t8zs78BrotqEJEh\nBvJ5evsG6OkboLc/vO3pGwjb+nNs2d7Db9/ZQeuOHtZv2TnseCccM4G5Jx5BU2OadE2K+roU2XEN\n1KVTo/BupFoVs8ayFrgUuD96/HF331Tw+h7CAFju7jmg3cxeB2YDc4Hbo76PAf/LzDJA2t1bovYn\ngPmEaz/LANx9nZmlzGxSNMbigjEWo2CRQ0RPX45tHb10dPXT2d3P25s7ad3eTU//YFgMsLmti46u\nflKpBLmB4Y4TCjXUpThu6jjeN21cuOZRmyRdkyJdkyQzJs2R2UamHzme3q7eEX6HcjAaNljc/WEz\nO7rg8SYAM/swsACYB1wA7Ch4WScwDsgUtHcUtLUX9O0AZgLdwNYh7XsbQ+SgFgQBa9/ZwZIf/Yru\n3oG99kskoK42Rbo2RX06xZGTGqmrDdcs6mtT1KVT4eN0DfXpFA11NdhR4xnXmB52p3FTY5pWBYuU\noKSd99FmrS8BF0X7TNqBpoIuGWAbYYBkCtq2E4bDnvr2FfQl6lM4Rm/BGEXJZjPDdyqzaqgRVGdc\nNrd18YtVG+jrz9M/MEB/tG+jP5dndUsbq1va6I02Y+Xz4drHH8w+kiMnNdLUWMf4sWlmTB3H+LF1\n1NfVkK5JDhsQB6LSl+cg1VlZ9jtYzOxKwp30Z7v74If8i8Dfm1kaaABmAauA54GLgJXRbbO7d5hZ\nr5nNAFqA84FbgAHgdjO7A5gOJNy9zcyei177PeBCoLnYWiv9SJFqOZpFdR64fBDwk+bf8dPnW/bZ\nb9zYNNnx9aRrU9TVJDn+qAl89MPHvKdfX3cffd19I1NspJKXZyHVGZ+4gm+/gsXMksDXgTeBh80s\nAJ5x9781s7uA5YTn3Sx09z4zWwrcZ2bNhGscV0RDXUd4pFcSWObuK6Lxm4EXojEWRH0XR2NcA2wp\nGEOk4rTv7KOto4funhxdvQN09fbT1ZPj6VfWs6mtC4AzTjyCmYdnqK1JUptKUlOTpCaVoLG+luOm\njTugQ29FKkEiGO6A9OoVVMO3g0qvEVTnvgRBwNOvrOeZV95h49Yu+nL7PhnwC5d9kHm/f7SWZ4xU\nZ3yy2Uws32p0gqRICbZs7+aBJ9eweVs4HQnA0VMyjBub5vDDxtBQVxP9n2JMXQ11tSmmZscyIVNX\n5spFRp6CRWQvgiCgdUcP/tY2WjZ00LKxna3tvXT15HabpmTKhAau/aMPMOOIpn2MJnLoULCIFAiC\ngIeb3+A3b23n7c2d9PS9e6hvKplg4rh6xo9NU1cbniD43y89kXGN6XKVK1KRFCxyyOvq6eft1p28\nsnYLzb9az86eHABTDhvDSVPGMvOIJo4/ajxTJ42ltkaXMBIZjoJFDmk/enot/7Hy7d12us84oolL\n583g92ZMLGNlItVLwSKHpOde28C9/2/1rll6Tzk+yweOmcAJxxzGlMPGlLc4kSqnYJFDSj4fcNuD\nv2Tt2+/OQPTlT5zKsVM1U5BIXBQsckgIgoCXvJWl/3cVQRBOwnjDx05i1tETyl2ayEFHwSIHtXwQ\nsOat7fz4md/y2/Xh3KeTxzfwmT8+kWnZsWWuTuTgpGCRg9bPXn6Hp375Nu+0htcdmZpt5Mr5x2NH\naS1FZCQpWOSg05/L8+2f/poVv9kMhFc7/JBlOfvkqSM6E7CIhBQsclB5Ze0W7n10NZ3d/QDM/9B0\nLj/vfWWuSuTQomCRg8LOnn5++NCvePyFFgDO+MDhnHvqNGYeqWlWREabgkWqXl//ADcsCS/Tk65N\ncuOfzNZ+FJEyUrBIVevq6eczUagc1lTPzZ/8EE1jNHeXSDkpWKRqtbX38PlvPQ9ATSrJ7Z+ZSyq/\n7+uhiMjIU7BI1QmCgF+s3sQ/P/JrAGYckeFLV57K4RMbK/5CSiKHAgWLVJXcQJ5/euQ/eclbAciO\nr+eLf3YqNSnNOixSKRQsUjVaNrbzjR+/xraOXjJjavn0xSdw0rGagVik0ihYpOK9/Horz7+2kZfW\ntO5qW/TpObrAlkiFUrBIRVv7zg6+8ePXAJh5ZBNnfOBw5s0+gtqaVJkrE5G9UbBIRRrI57n/CefZ\nX20A4I/PmslFpx+tKVlEqoCCRSpOEATc/J0X2bC1C4CjJo9VqIhUEQWLVIzBa6Z86yerdrX99Z+c\nxEnHTipjVSKyv4oKFjObA9zm7ueY2bHAd4E8sMrdF0R9rgGuBfqBxe7+qJnVAw8Ak4F24Cp332pm\npwNLor5PuvuiaIybgYuj9hvdfYWZTQQeBOqB9cDV7t4Tz9uXSpEbyPPVH7y868qOxxye4ZMXzuKo\nKZkyVyYi+2vYg//N7AvAPUBd1HQnsNDdzwKSZnaJmU0BbgDOAC4AbjWzWuB64FV3nwfcD9wUjbEU\nuMzdzwTmmNlsMzsZmOfuc4DLgbujvjcD349+3ivAdQf8rqWi9PYPcO3Xnmbt2zuoSSX47H87iZs/\n+fsKFZEqVcxZZWuBSwsen+ruzdH9x4D5wGnAcnfPuXs78DowG5gLPF7Q91wzywBpd2+J2p+IxpgL\nLANw93VAyswm7WmM/X2TUrmCIOCrD7686/FXrj2dDx6nTV8i1WzYYHH3h4FcQVPhHtQOoAnIADsK\n2juBcUPaOwra2oeMMbTv3toH2+Qg8chzLfxuQ/jncOtfns6kcQ1lrkhEDlQpO+8LZ/nLANsJg6Jp\nSPu2qD0zpG/HXvr2FfQl6lM4Rm/BGEXJZit/U0o11AgjV+eLq8OrPH7hylP5veOnHPB4h/ryjJvq\njFe11HmgSgmWX5rZPHd/FrgQeApYASw2szTQAMwCVgHPAxcBK6PbZnfvMLNeM5sBtADnA7cAA8Dt\nZnYHMB1IuHubmT0XvfZ70c9rpkiVPiFhNpup+Bph5Op8cfUmNmwNr0c/a2rTAf+MQ315xk11xqsa\n6owr+EoJls8D90Q751cDD7l7YGZ3AcsJN5UtdPc+M1sK3GdmzYRrHFdEY1xHeKRXEljm7isAon4v\nRGMsiPoujsa4BthSMIZUuQefXAOga9GLHGQSQRCUu4aRElTDt4NKrxFGrs5P3fYUk8bVc/t1Z8QS\nLIf68oyb6oxXNdSZzWZi+YanucalLN7e3AnAlAkNWlsROcgoWKQslq1cB8Dx08eXuRIRiZuCRUbd\nq7/dwvJXw8klPzRrcpmrEZG4KVhk1C350asAzDpqPEdMbCxzNSISNwWLjKoXV2/adf9/XPbBMlYi\nIiNFwSKjprO7n/sedwDO+9A0Ukn9+YkcjPQvW0bNr1va6O7NcfL7JnH5ue8rdzkiMkIULDJq8tE5\nUyfOnKhDjEUOYgoWGTUDAwftybgiUkDBIqPm355rAUDxInJwU7DIqPi351vYvL0bQNdbETnIKVhk\nxK1+cxsPP/sGAAuvPJUJmbphXiEi1UzBIiOqs7ufr/0gvELkeadO47hpuk6byMFOwSIjJh8EfPbr\n4eVzJmTquGL+8WWuSERGg4JFRsxjP39z1/2/+/ScMlYiIqNJwSIjpqsnB8CfnH0sY+pLuaaciFQj\nBYuMmMd+8RYAdtSEMlciIqNJwSIjoq29Z9f9qVnNYCxyKFGwyIh4btVGIJwav642VeZqRGQ0KVgk\ndgP5/K7zVk4+PlvmakRktClYJHZPv7x+1/1zT51WxkpEpBwULDJiPjZvJknNYixyyFGwyIiZctiY\ncpcgImWgYBERkViVdNaamdUA9wHHADngGmAA+C6QB1a5+4Ko7zXAtUA/sNjdHzWzeuABYDLQDlzl\n7lvN7HRgSdT3SXdfFI1xM3Bx1H6ju68o6d2KiMiIK3WN5SIg5e5/APwd8BXgTmChu58FJM3sEjOb\nAtwAnAFcANxqZrXA9cCr7j4PuB+4KRp3KXCZu58JzDGz2WZ2MjDP3ecAlwN3l1iziIiMglKDZQ1Q\nY2YJYBzhmsQp7t4cPf8YMB84DVju7jl3bwdeB2YDc4HHC/qea2YZIO3uLVH7E9EYc4FlAO6+DkiZ\n2cQS6xYRkRFW6gROncAM4DfAROCjwJkFz3cATUAG2DHkdeOGtHcUtLUPGWMm0A1s3cMYhW0iIlIh\nSg2WG4HH3f3LZjYVeBpIFzyfAbYTBkXTkPZtUXtmSN+OvfTtK+hb2H9Y2Wxm+E5lVg01wv7VmRkb\nXsirqal+1N/fwbg8y0l1xqta6jxQpQZLG+HmLwg/5GuAl83sLHd/BrgQeApYASw2szTQAMwCVgHP\nE+6nWRndNrt7h5n1mtkMoAU4H7iF8KCA283sDmA6kHD3tmKKbG3tKPHtjY5sNlPxNcL+19nR2QtA\ne3vPqL6/g3V5lovqjFc11BlX8JUaLEuAe83sWaAW+CLwEvDtaOf8auAhdw/M7C5gOZAg3LnfZ2ZL\ngfvMrBnoBa6Ixr0OeJBw38+ywaO/on4vRGMsKLFmEREZBSUFi7vvBD6+h6fO3kPf7wDfGdLWDfzp\nHvq+SHgE2dD2RcCiUmqV0ZcPgnKXICJlpBMkJXYveSsASc3mInJIUrBIrHIDedas204COHGmjgoX\nORQpWCRWK3+zGYBZR08greuwiBySFCwSq/uXrQHg92YeVuZKRKRcFCwSmzc3dtDdmwPgIyfrOiwi\nhyoFi8TmR0+vBaCpMU1dWpvBRA5VChaJRT4f8OuWbQB8+ROnlrkaESknBYvE4s1N4RnFTY1psuMb\nylyNiJSTgkVi8au1WwC44LSjylyJiJSbgkUOWG//AI881wLoaDARUbBIDJ55+R0AalJJpmXHlrka\nESk3BYscsEd//iYAH5s3s8yViEglULDIARnI5+noCq+gcP5p08tcjYhUAgWLHJDBQ4xPft8kEgnN\nOikiChY5QINzg51yfLbMlYhIpVCwyAFZs247qWSCOSdMKXcpIlIhFCxSss7ufjZt62byhAZqUvpT\nEpGQPg2kZIMnRR5+2JgyVyIilUTBIiV7c2M4jcsxh2fKXImIVBIFi5TsuVUbATh++vgyVyIilUTB\nIiXZ2dO/69orx00bV+ZqRKSSKFikJK+9sRWA2cdOJJXUn5GIvEufCFKSTW3dAJys81dEZAgFi5Sk\nLzcA6IgwEXmvmlJfaGZfBP4IqAW+BTwLfBfIA6vcfUHU7xrgWqAfWOzuj5pZPfAAMBloB65y961m\ndjqwJOr7pLsvisa4Gbg4ar/R3VeUWrfEIwjC21RS07iIyO5KWmMxs7OAM9z9w8DZwFHAncBCdz8L\nSJrZJWY2BbgBOAO4ALjVzGqB64FX3X0ecD9wUzT0UuAydz8TmGNms83sZGCeu88BLgfuLvG9Sow6\no4knxzbUlrkSEak0pW4KOx9YZWY/AR4Bfgqc4u7N0fOPAfOB04Dl7p5z93bgdWA2MBd4vKDvuWaW\nAdLu3hK1PxGNMRdYBuDu64CUmU0ssW6Jyevv7ABg/Ni6MlciIpWm1E1hkwjXUv4QmEkYLoUh1QE0\nARlgR0F7JzBuSHtHQVv7kDFmAt3A1j2MUdgmo6irJ8emti4A0rXaTSciuys1WLYCq909B6wxsx5g\nWsHzGWA7YVA0DWnfFrVnhvTt2EvfvoK+hf2Hlc1W/hnh1VAj7F7nUyvfAmDGkU1Mnty0t5eURTUu\nz0qmOuNVLXUeqFKDZTnwWeD/mNmRQCPwH2Z2lrs/A1wIPAWsABabWRpoAGYBq4DngYuAldFts7t3\nmFmvmc0AWgg3t90CDAC3m9kdwHQg4e5txRTZ2tpR4tsbHdlspuJrhPfW+YqHU+WffsKUiqq/Wpdn\npVKd8aqGOuMKvpKCJTqy60wzexFIEO6MbwG+He2cXw085O6Bmd1FGEQJwp37fWa2FLjPzJqBXuCK\naOjrgAcJN6stGzz6K+r3QjTGgtLeqsRlR2cfAO8/ekKZKxGRSpQIBo8bPfgE1fDtoNJrhN3rDIKA\nG7+xnNxAwJLPzq2o6fKrcXlWMtUZr2qoM5vNxHL+QOV8KkhV2LC1i/aufk48dmJFhYqIVA59Msh+\nWRsdZny8Jp4Ukb1QsMh++fl/hlPlj6nXiZEismcKFtkv26Md97O0415E9kLBIkXr7s2xsa2LiU11\njGtMl7scEalQChYpWkt0KeL6upLnLhWRQ4CCRYrW1t4DwKm6BouI7IOCRYq2Y2e4fyWR0FT5IrJ3\nChYp2mCczDji0JjvSERKo2CRovX0hVeN1BqLiOyLgkWKtup34ZUKpk5qLHMlIlLJFCxStMF9LJkx\nOjlSRPZOwSJFSwATMnXU1qTKXYqIVDAFi+yXVFL7V0Rk3xQsUrTcwEF7iQURiZGCRYrS3tXHjp19\nTJnQUO5SRKTCKVikKC0b2gE4rKm+zJWISKVTsEhR1m3uBOCoKTo5UkT2TcEiRRk81PiIiWPKXImI\nVDoFixRly/ZwAspJ47WPRUT2TcEiRcnl8wCM13VYRGQYChYREYmVgkWK0t+fL3cJIlIlFCxSlLc2\nh1ePTOrMexEZxgFdY9bMJgMrgfOAAeC7QB5Y5e4Loj7XANcC/cBid3/UzOqBB4DJQDtwlbtvNbPT\ngSVR3yfdfVE0xs3AxVH7je6+4kDqlv2XTCSYkKmjJqXvIiKybyV/SphZDfCPQFfUdCew0N3PApJm\ndomZTQFuAM4ALgBuNbNa4HrgVXefB9wP3BSNsRS4zN3PBOaY2WwzOxmY5+5zgMuBu0utWUqTzwfs\n7Mkxpl7XuheR4R3I189/IAyC9YQT357i7s3Rc48B84HTgOXunnP3duB1YDYwF3i8oO+5ZpYB0u7e\nErU/EY0xF1gG4O7rgJSZTTyAumU/bWoLvztM1Fn3IlKEkoLFzD4JbHb3J3n3irWFY3UATUAG2FHQ\n3gmMG9LeUdDWPmSMoX0Lx5BR1qRDjUWkCKVu27gayJvZfMI1kO8B2YLnM8B2wqBoGtK+LWrPDOnb\nsZe+fQV9C/sPK5ut/OlHqqHG9a3hdC4N9bUVX2+l1zdIdcZLdVaWkoIl2o8CgJk9BVwHfM3M5rn7\ns8CFwFPACmCxmaWBBmAWsAp4HriIcMf/RUCzu3eYWa+ZzQBagPOBWwgPCrjdzO4ApgMJd28rps7W\n1o5S3t6oyWYzFV8jwNubdwKQTiUqut5qWZ6qM16qMz5xBV+ce2M/D9wT7ZxfDTzk7oGZ3QUsJ9xk\nttDd+8xsKXCfmTUDvcAV0RjXAQ8SblZbNnj0V9TvhWiMBTHWLEXo6csBmtlYRIpzwMHi7h8peHj2\nHp7/DvCdIW3dwJ/uoe+LhEeQDW1fBCw60FpFRGTk6aQEERGJlYJFhrVhy85ylyAiVUTBIsPq6OoH\noCGdKnMlIlINFCxStCMmNpa7BBGpAgoWERGJlYJFRERipWAREZFYKVhkWPkgKHcJIlJFFCwyrLc2\nhnODTmiqK3MlIlINFCwyrP5cnkQCmsZodmMRGZ6CRUREYqVgkWH15/LlLkFEqoiCRYa1fstO0jU6\n615EiqNgkWH19g0w5bCGcpchIlVCwSJFqU3pT0VEiqNPCxERiZWCRfapo6uP3ECeGq2xiEiR9Gkh\n+7S9sw+AwyeOKXMlIlItFCxSFK2xiEix9GkhIiKxUrCIiEisFCwiIhIrBYuIiMRKwSIiIrGqKeVF\nZlYD3AscA6SBxcCvge8CeWCVuy+I+l4DXAv0A4vd/VEzqwceACYD7cBV7r7VzE4HlkR9n3T3RdEY\nNwMXR+03uvuKkt6tiIiMuFLXWK4Etrj7POAC4JvAncBCdz8LSJrZJWY2BbgBOCPqd6uZ1QLXA69G\nr78fuCkadylwmbufCcwxs9lmdjIwz93nAJcDd5dYs4iIjIJSg+WHvBsGKSAHnOLuzVHbY8B84DRg\nubvn3L0deB2YDcwFHi/oe66ZZYC0u7dE7U9EY8wFlgG4+zogZWYTS6xbRERGWEnB4u5d7r4zCoMf\nAV8GEgVdOoAmIAPsKGjvBMYNae8oaGsfMsbQvoVjiIhIBSppHwuAmU0H/hX4prv/i5l9teDpDLCd\nMCiahrRvi9ozQ/p27KVvX0Hfwv7DymYzw3cqs0qvsbM/vMhXQ0NtxdcKlb88B6nOeKnOylLqzvsp\nhJuqFrj7z6Lml81snrs/C1wIPAWsABabWRpoAGYBq4DngYuAldFts7t3mFmvmc0AWoDzgVuAAeB2\nM7sDmA4k3L2tmDpbWztKeXujJpvNVHyN27Z1AdDd3V/xtVbD8gTVGTfVGZ+4gq/UNZYvAeOBm6Ij\ntgLgr4BvRDvnVwMPuXtgZncBywk3lS109z4zWwrcZ2bNQC9wRTTudcCDhJvolg0e/RX1eyEaY0GJ\nNYuIyChIBEFQ7hpGSlAN3w4qvcZ1mzv53/e+yLmnTuPP5h9f7nL2qRqWJ6jOuKnO+GSzmcTwvYan\nEyRFRCRWChYREYmVgkVERGKlYBERkVgpWEREJFYKFhERiZWCRfZpc3SCpIhIsRQssk87e3IA1NWm\nylyJiFQLBYsUZeqkxnKXICJVQsEiIiKxUrCIiEisFCwiIhIrBYuIiMRKwSIiIrFSsIiISKwULCIi\nEisFi4iIxErBIiIisVKwiIhIrBQssk9d0VxhIiLFUrDIXnV09fHDn60FoLZGfyoiUpyachcglWnD\n1p18/aFXdz2efdykMlYjItVEwSIABEFAbiDgdxvaWf7aBp57dQNB9Nw9C88jlc+XtT4RqR5VESxm\nlgC+BcwGeoC/cPc3yltV9cnnA3728jts2tZFfy7PG+vb2bKjm/5cQG5g9+CYOqmR//L70znp2Ikc\nPrGR1taOMlUtItWmKoIF+K9Anbt/2MzmAHdGbbIfHm5+g0dfeHPX42QiwZGTxlBbk6K2JkltKsG4\nsXXMOWEK7z96AjUp7VcRkf1XLcEyF3gcwN1/YWYfKnM9VeeXa1p59IU3mTy+gb+85APU1aZoakwz\ntqG23KWJyEGmWoKlCdhR8DhnZkl3H7EN/xvbuujqyREQEP0X3QYE0c6HILoT7Ho+IGD3voOvDaLG\noGCspk2d7NjRTfRwV5+BfMDAQED/QJ6BgTz9A+Gmqnf/D8jl8uTy0W1he8H9/oL7W3f0kK5JsuBj\nJzJ98thuy4zZAAAFbUlEQVSRWmwiIlUTLO1ApuDxiIbKmnXbue37vxyp4UdFMpGgpiZBTTJJTU2S\nw5rq+NNzjlOoiMiISwx+665kZvYx4A/d/VNmdjpwk7tfXO66RETkvapljeVhYL6ZPRc9vrqcxYiI\nyN5VxRqLiIhUDx1PKiIisVKwiIhIrBQsIiISKwWLiIjEqlqOChuWmc0Cfg5Mdve+6LDkJUA/8KS7\nL4r63QxcHLXf6O4rRqm+JuABwpM9a4HPRbMIVFSdQ2quqDnazKwGuBc4BkgDi4FfA98F8sAqd18Q\n9b0GuJZw+S1290fLUO9kYCVwHjBQiXWa2ReBPyL8m/wW8Gyl1Rn93u8j/L3ngGuooOUZTTN1m7uf\nY2bHFluXmdUTfiZMJjxX7yp33zpKdX4QuItwefYCf+7urXHVeVCssZhZBvgHwg+/QUuBy9z9TGCO\nmc02s5OBee4+B7gcuHsUy/wc8O/ufjbh4dLfqtA6C+2aow34EuEcbeV0JbDF3ecBFwDfJKxpobuf\nBSTN7BIzmwLcAJwR9bvVzEZ17prow/Afga6oqeLqNLOzgDOi3+/ZwFGVWCdwEZBy9z8A/g74SqXU\naWZfAO4B6qKm/anreuDV6O/5fuCmUaxzCbDA3T9CeDrH38RZ50ERLMA/E37wdcGuoEm7e0v0/BPA\nfMI5x5YBuPs6IGVmE0epxjuBf4ru1wLdFVpnod3maAPKPUfbD3n3jzpF+G3rFHdvjtoeI1x+pwHL\n3T3n7u3A68BJo1zrPxB+aVgPJCq0zvOBVWb2E+AR4KcVWucaoCZagx5H+G26UupcC1xa8PjUIuua\nTcG/r6jveaNY58fd/bXofg3hl/LY6qyqTWFm9ingRqDw5Ju3gB+4+2vRHx6Em5vaC/p0ADOBbqBw\nFa6T8A811tXPIXUmotur3f0lMzucMPU/W+46izDqc7Tti7sXfnH4EfBlwg/wQR2ENWfYve7B5Tcq\nzOyTwGZ3f9LMFkbNhV/iKqJOYBLhWsofEv7dPUJl1tkJzAB+A0wEPgqcWfB82ep094fN7OiCpkTB\n/eHqKmwf7Dsqdbr7JgAz+zCwABjcChBLnVUVLO5+L+E29l3MbA3waTP7C+Bwwm/6H2X3N58BtgF9\n7D7nWAbYPhp1RrWeCDwI/A93Xx59QJatziKM6hxtxTCz6cC/At90938xs68WPD24nNp573IdzeV3\nNZA3s/mE3/i+B2T3UE+569wKrHb3HLDGzHqAaXuop9x13gg87u5fNrOpwNOE+9iG1lPuOiHctzL0\n5++prm3s/u9r1Gs1s48Tbum5yN23mllsdVb9pjB3P97dP+Lu5wAbgfnu3gH0mtmMaC3mfKAZeB44\n38wSZnYUkHD3ttGo08xOINyUc4W7D27mqrg6h3iOcPs20UEGr+27+8iKtgE/AfxPd78van7ZzOZF\n9y8kXH4rgLlmljazccAsYNVo1enuZ7n7OdHf5CvAJ4DHKq1OYDnht1TM7EigEfiPaN9LJdXZxrvf\nmLcTfiF+uQLrBPjlfvyenyf69xXdNg8dbKSY2ZWEaypnu/vgRZpejKvOqlpjKcLgpieA6wjXDpLA\nssGjqsysGXgh6rdgFGv7CuGOs69HIbLd3S8l3DFWSXUWqrQ52r4EjAduio6aC4C/Ar4R7WRcDTzk\n7oGZ3UX4wZkg3JnaV66iI58H7qmkOqMjfs40sxejn3890AJ8u5LqJNzRfK+ZPUu4f/KLwEsVWCfs\nx+/ZzJYC90X/1nuBK0ajQDNLAl8H3gQeNrMAeMbd/zauOjVXmIiIxKrqN4WJiEhlUbCIiEisFCwi\nIhIrBYuIiMRKwSIiIrFSsIiISKwULCIiEisFi4iIxOr/A0/teZ21UxPbAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x124f22110>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(pc.diff_acc_to_purchase, np.cumsum(pc.total_spent))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "112.99273255813954"
      ]
     },
     "execution_count": 299,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pc.diff_acc_to_purchase.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 301,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x12867bb50>"
      ]
     },
     "execution_count": 301,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAECCAYAAAD3vwBsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12867b9d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pc.diff_acc_to_purchase.hist(bins=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb8AAAERCAYAAADrBansAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12867bd10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "with plt.style.context('fivethirtyeight'):\n",
    "    plt.plot(pc.diff_acc_to_purchase, np.cumsum(pc.total_spent))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 305,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAakAAAERCAYAAADBtVhDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKQCAsSgpAICxKCkAgLH+PzymRinTuZZvAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1274f4090>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "with plt.style.context('fivethirtyeight'):\n",
    "    pc.diff_acc_to_purchase.hist(bins=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tmp = pc[['diff_acc_to_purchase', 'total_spent']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 341,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def negative_to_zero(x):\n",
    "    if x < 0:\n",
    "        return 0\n",
    "    else:\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 342,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tmp.diff_acc_to_purchase = tmp.diff_acc_to_purchase.apply(lambda x: negative_to_zero(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 366,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average time since registration for an user to make the first purchase: 103.324216524 days\n"
     ]
    }
   ],
   "source": [
    "tmp = tmp.groupby(tmp.index).agg({'diff_acc_to_purchase': np.min})\n",
    "print \"Average time since registration for an user to make the first purchase:\", tmp.diff_acc_to_purchase.mean(), \"days\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 883,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ8AAAFHCAYAAAB+j0HWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xm8VVX9//EXkqjEJHodch66nwbTRCvryw9zyhxKlLJJ\nMcskJY00LS0NzUxFGzSHciTRMgfUEiunHDC1QHH+OIuFJgLCBRzh/v74rMPdHO6wz7337HvP4f18\nPHhwz157n73WOfvsz17D3qtPc3MzIiIiRVqlpzMgIiIrHwUfEREpnIKPiIgUTsFHREQKp+AjIiKF\nU/AREZHCKfiIiEjhFHxERKRwCj4iIlI4BR8RESmcgo+IiBROwUdERAqn4CMiIoVT8BERkcIp+IiI\nSOEUfEREpHAKPiIiUjgFHxERKZyCj4iIFE7BR0RECqfgIyIihVPwERGRwin4iIhI4RR8RESkcO/p\n6Qx0hZn9BPhJetkMzAeeAf4OnOPu/+upvFUilWOsu6/T03kBMLMdgTsyixYATwI/d/cbqrTPTYDn\ngb3dfUqO9VcFjgcmu/vD3ZiPLwL93X1i2fI7gNnuvn937auDfBwD3O/udxWxv66o9Lur8L3Lj8WS\nd929X1rnbuAld/9qF/f1CeAz7v7THOtOAkr7Wwr8B7gVOM7dZ3clH3l1V7l7Sj3UfF4HPgF8EvgS\ncC1wIPCImW3bkxmrwIXA7j2diTLNwFeAHYAvA68B15nZ8Crt7+W0r3tyrt+PuPD4aDfnY3/goFaW\nHwYc1837as+xwKcL3F9XNVf5vUvHYunf/2XSvwX8uBv2swNwQgV5epQ49wwHTgH2ASZ3Qz5WCjVd\n80nedfd/ZV7fYmbnA3cDfzSzD7h7NX8YXebus4BZPZ2PVjzi7o8DmNmdwEvAAeQPELm5+9vAAxVs\n0ifvima2uru/WXmuWrj7k13ZfiWQ+/vopGXHYrk8342Z9UvHWHsqLcPCzLnnPjN7G7jUzLbuSm28\nO47XWlAPwWcF7r7AzI4FpgC7Ec1wmNnPgb2AzYga053A0aXmOTM7HfiCu2+RfT8z+zrwW+B97j7H\nzL4JHJXeZxFxBXS4uz/RWn7MbDBwFrAHMBR4Ffiru49J6eOJZreG9LrU1LATMDZt9ypwprufX/be\nI4DxwMeAJcB04HvuPiOlbwRMSJ/D6kRQPtLdn8r9gQLu/oaZPQNsVLb/PsAPgG+mtBeBn7n778vW\n+ylwaMrD1UQTxZXApu4+s7WmGzP7PHAi8AHgbeAp4Bh3v5toCmwGLjOzy9LfmxEnkOeJILk78Hng\nX8BnzOzAlIcPpfUeSu83Le3vUmAU0GxmS9N7nuTuJ5vZP4BXs81uZrYzcCqwDdHkey1wrLsvSum5\nv8eyz+p54jgZn46NZmAnd7/LzNYATge+CAwBHgF+5O63tPV+6T2XEsfshsDX03ue5u5nmdlB6XMe\nClwHHFY6UZvZesDPiFrY+sQFyJ+Ak939nXb2txNwA3C2u/84LeuWY7GVfS3X/GRmpxDH4/7AL4EP\nA183s+uIz+4LwDrAHOCfab2DgV/Q8t0D3Orun6kgK9OJ42oz4GEzewm43N2Pz+T1EOB3wOru/raZ\n7QLcQnwm44jPeRJwmJmtQtS2DwI2BmYT541vlZX/AOIc0EBcGH7T3V/JpJ8BfJaW894dxHlvdmad\nfYnaoxG/tSeJ38bUlF7Ky8HEMfQCcIq7T6rg81lOPTS7teUfxMl4h8yydYFSAPou8WXclkm/BNg0\nndCzvg7ckALPCOB8YCLxhR4M3AsMbicvvwQ+lfb5GeJLzNbGmmm92eJ3xAlyJHHA/MbMti8lmtmn\niZP4W8Bo4kd0N7BBSl8TmAq8nzjpfhF4L1E7XK2d/K4gBZkNgefKkn5D9L1cAOxJnLwuMbM9M9t+\nL5X5POLk/gZxEigvc3Nmm81pCVJ7E+3rfyZOkAA7Ez/0k4nv+JNE013JBCJAfYEIEBDf9yTic/gK\nMBO4y8w2TeknE5/zg7Q05V5UnreUvw8DNxPBZD/i5P3VlOdy7X6PrRiZ8n5RpmzTU9pFxMnop2m9\nmcBNZvapdt6v5Cji+/8ycAUwIV1wjQaOIL6jrxEnwZK1gXnA0UQwP4P4PZzd1k7MbHfgL0RwKwWe\nrh6Lfc0s+6+9WkozMJD4PZ9PBP1pRJPaF4njdddUzibiPHgD8CvinPEJ4nM/Ike+sjZL+365nXXa\n+q1fAvwb+BxwWWbZCcR3tRfxHby3bLvhxOc5DhgDbE/8FrPWoeW8N474Dm4tJZpZI/BH4G9pna8S\nF+5rZt7jfKIp+Fzid34DMNHMKgnOy6nLmg+Au79lZq8RAae07Bulv1Mkvx94ycyGu/s97u5mdi8R\nUO5K620O/D/iBAhRw5jh7mdkdveXDrLzMeBcd78ms+zKHMW40t1PTfm4k7iK3484SCEOqAfdfY/M\nNn/P/H0UsAaws7vPT+9zL3HV8g3igGpPXzPrS5zwjwX6kznpmNkWwLeBgzJXQLeb2fuI/pgp6XM+\nBjjP3U9K69yaPtcNy/aXPaFsCyxw9x9mlv0183epueM5d1/WXGdmpT//6e7LnTzc/eTMen2IH+An\niFrSKe7+vJnNBfqUNeW25gTic9yn1KxrZvOAq8zsE+5+f2bdjr7H5bj7DDN7F/hPWdk+QASOZZ+3\nmf0deDjlZ4/W3i/jKXc/LG13G3Gxcgiwcaa2thOwLxFkcPdHge9n8nAvsBi42MyOcPd3sztItdWr\ngOPd/ZeZpK4ci32AGZnXzURt7MR2tlkDOMLdlx0zZvYxYFLZ1XrpYuE1M3sxlbmj736Z9PtYBfgI\n8Xt8gbjQqNQfMr+P0sXNaKIW+tvMen8q2+69wJ7uvjBttwFwupm9p/TduPvXM++7CnHcPW9mO7j7\nfcRvbW62hkbmt5aC07eAr7r7H9Pi281sQ+J3nj3n5Fa3wSdZ7urIzPYgqpYfBgalxc1AIy39GBcD\nZ5vZWHdfTFzlvUJcFUAcWKeb2S+IzsX72mt+yGxzbKrO3+ruT+fIezNRHQfA3d81s6dJJ2wz6w98\nnPavzkpV+oXpRwKwkLgKbO/KG1b8wS8BRpXlfZe0/PrM+wPcDnw5neA3BtYjai1ZNxI1x7Y8AgxO\nTWpXAFPT95HXCqOuzOyDRC3ok8TVILR8/5X6GHB1WX/itcC7xNVoKfi0+z12Yp8Ayy5i3L3ZzK4m\nAnxHbi/b7nlgUSnwJM8Qn88yZjaOOPlsRjSXQZRrY5avCX+BuGoe5+7lV99dORYhAmV2Xx31kS6h\n5Tdb8hBwSLoo/VsKrF2xA1D67TcTFwGjc/QtlWtmxeN1p7R84oqrL+f+UuBJHid+u+8jasWY2d5E\nbe9DrHjeuy/le20zu4S4KJ7q7m9k3nNXoinuxlZ+5+fkKWBr6rbZLVXl1wJK/TkfI6qKM4kr3R2I\nq94+tPygIK4smomDHeLqY6K7LwVw99uImtH/I5pQXjOz36S2+LaMJQLVCYCb2VNm9qUcxXi97PXb\nmbyumfL+Cm1bmxgB+E7m39tEu/JGbW+2zP7EiWE/wIn+lfXK3v89RBNRdh+XAn2JPoJ1ic+zfPhp\nu8NRUz/APsQJ7ybic77CzNbOkW9I33uJmQ0grtA2AL5HBIjtiR/e6its3bH1y/eRjpE5tDQNlrT3\nPVa6z4WtdEb/D+hvMfy8Pa3lo928pSbTCURg/TwRAMem5PIyfI4o//Wt7Lsrx2Iz8Li7T8/8a++4\nB5jTykCj8UQN6zvADDObaWZjV9gyv0eA7Yiaw7ruvq27T+9gm7aU3xayFjA/x8CD1r4/SN+Nme1A\nnHueI5pUdyC6AJad91Jf9UhgSyIIvmZml5tZ6ThemxhdupDlv78LgX5m1qlbROq55rMzUb570+uR\nRIfxV0ormNnG5Ru5+2Iz+yPRQTmT+GFcVrbO5cDlZrYWcWL+FXECPp5WuPsCoq11nJltRTRhXWFm\nM7owimoecX/B+u2sMxd4jOjLKG8jb+rg/Us/+MeB6Wb2MHFVdQItJ5+5xEH4KVpvx34VWDXtu6Es\nrfz1Ctz9ZuBmMxtItEX/mmj2y3NfQ3l+PklcDe6crb1ZDAbpjJdpqT2V3msV4qQxt5PvmWefA2zF\n0VDrAotz1MA74wtEDW9ZE1dqEmrNEUTz2i1mNsLd52XSunIsdsYKx6O7v0U01Z1oZlsChwPnmNkT\n7n57+fo5LHL3B9tJf5M4aWet2dqKreR3DlHz7+rIt32BWe5+QGlBavJejrvfRPQdDiQuIn6V/o0m\nvru3WH54e3leK1aXwcfMhhAd2k/RMqBgDVqqyCUH0PpJ82KiOjqeaFZrdTSOu88BLjSzUUSVtkPu\n/qjFSLwDiFFcnQo+KUjeTxwc57ax2m1EB+vj6YfXae7+nJldBBxsZj9x99eIandfYEiqEa7AYsTP\nK0QtJjsia58K9t1EDJv/NC0DSJa7wsuhVDNd1iSSOuk3Zfm+l7y1kvuBfc3s+MwV9iji8+iOoeit\n5aPUF/EFYuAEmdd3d8M+W7MGceLJOqC1FYkLsN2J/tK/m9lOmSahbjsWu4O7P2NmRxO1oA8Rx/Lb\nRD/nKqWWji76D/DBsmV57+e7jQjSo4kBK521BpljPmnrvFf6rV1pMZKzdA/d7UQQHejud3YhL8up\nh+DzHos7kyFGuGxH3BC4BrB75sRwC/BdM/sl0f/wKdr4Ebn7A2b2GBHpD82mWQx9HUqMpnsNGAaM\nIIYbt8piKOhkYkh2c3rPhbR/X0ueew5+SFxl3kwcoIuIK/x/eQxX/gVR1b7DzM4B/ktcJe8I3O3u\nV1W4/zOItv8jgJ+4+1Nm9luik/0M4iS+OtGn9n53P9Tdl5rZBGJk1WvEiKfPA1ul92z1R25mh6ay\n/JVo328kTl6XAbj7O6nPYv/0Xb3J8n1U5e5Ln89FKa8bEZ2l/ylb70ng82a2T0qb5e6tjV46hRiB\ndoPFfWUbAacRQ2Gzgw06e//Lk8BeZvY34lh50t2fNLM/EKPlBgHPEseSESOdquEW4AgzeyDt72vA\nFm2t7O7zzGw3IhjeZGafTf0H3X0sVszMbiAuGh4iRlx+ifg9lp4iUboQPMpiaP38nP2zbZkMnGVm\nPyCOlS/Qev/iCuVz9ydSH8yvzWx94oJmKDDS3b/WwX6z73cLMNbMziKar4cT38MyZnYY0QT9N6J2\n/QGiRefClJfH04XnNem3M404v34Y2Kw0iKVS9dDnM5hoWptK9NeMAn4PfMTdl406SU04PyA+1BuI\nPpu92nnf64kDtPxH8S/iauZ84sQ4hjgRt9fxdi8xPPbq9H5Dgc963FzalrZujF223ON+l92IA+Fy\nYrjkCNIJNdXMdgCeIH78fyNqhIOIvo72tNZsMZO44j681Mfl7mOJppQDiYP7UmIo5l2Z7X5JdPQf\nRnSWD6Zl+POCNvb5MNHWfFbK9/HEvVbZ0W9j0jq3EIH8fe3k/VXix78u8d0embZ/pmzV84i+oYvT\ne2bvqch+9o8To8saiP6Qk4mBEV8se78Ov8c2HEMEy7+kfGyXlh9CdEKfkMqxEbCXu/+zg/dra4hv\nR/k4GfgDMbT7SiLItzbIJfvZvEIMMNgEuDaNvOrWY7GT7iGaoa4gPrutiZP5wynfd6S8jSMuVtpq\nUcjrfKJDfhzxGTbRctxntVW+bxEXOaXf1pms2ETZ7nfq7n8mfjv7E+e9T9Iycre03gzid/FL4tj/\nIfE7+FHmfcYQIwwPSnm5hBgw1Okad5/m5l5983+PSU1aT2SHKUr3SVdSu7j7Zj2dFxEpXj00u3Ur\nM9uOuGrbnrhSly5KHdRfImqAzUSN4SBi4IWIrIQUfFb0L2Ik2Q+7MGxSlreIaGseS9wU9yLxGJpf\ntruViNQtNbuJiEjh6mHAgYiI1BgFHxERKZyCj4iIFE7BR0RECqfgIyIihVPwERGRwin4iIhI4RR8\nRESkcAo+IiJSOAUfEREpnIKPiIgUTsFHREQKp+AjIiKFK3RKBTMbSswUuDsx3fJNwHfc/fWUPho4\nEVgPeAQYm53WwMy2J2YX3IqYWnm8u19RZBlERKTriq75XE7M57IFsBkxBfLlAGY2nJi6dQywJnAd\nMMXMBqT0QcAUYirqIcREbxeY2ScKLoOIiHRRYfP5mFl/Yv7xbdz90bRsBHAHMdf7KUAfdz8os80L\nwAnufrmZHQycmJ122cx+D7zj7t8spBAiItItiqz59En/svvsm/7fFtgGmFa2zUNpOcDWwINl6dMz\n6SIiUiMKCz7uvoio5Yw3s8Fm1gAcl5IHpn/zyzZ7HRiUWae9dBERqRGFDjgADiAGHDwBvAGcBewC\nvEY0yQ0uW38I8Ez6u4lonitPX9DRTi/5/VXNffr0yZ3JZuDze+zC2muvlXsbEZE6lP/EWaFCg4+7\nvwx8pfTazPYC3gTuA2YAw8o22Ra4Jv09A9inLH1YWt6uP0xdzOoDhubO5ztvLWa1Ve/hM7vsnHub\nntTQMJDZs5t6OhtVU8/lq+eygcpX6xoaBlbtvYseat0IzCaay7YHfgn83N0XmNmFwM1mNhGYCowD\n+gHXp80nA6eb2dHAOcAIYCSwa5FlEBGRrit6qPUI4DGiCW0ScLa7nwLg7lOBw4GLgHnAfsAe7r4w\npc8H9gT2T+kXAGPc/YGCyyAiIl1UdLPbRURwaSt9EhGU2kqfBui+HhGRGqfH64iISOEUfEREpHAK\nPiIiUjgFHxERKZyCj4iIFE7BR0RECqfgIyIihVPwERGRwin4iIhI4RR8RESkcAo+IiJSOAUfEREp\nnIKPiIgUTsFHREQKV/RkcusCZwM7AX2BB4Gj3P3hlD4aOBFYD3gEGOvu0zPbbw+cC2wFzALGu/sV\nRZZBRES6ruiaz/nAEGBLYF1gGvAXADMbDpwHjAHWBK4DppjZgJQ+CJgCXJ3e4zDgAjPT/D4iIjWm\n6OCzBXCNuy9w93eBi4ENzGwocAhwrbvf5u7vuPsE4E1g37TtKGCRu5+Z0m8lptY+tOAyiIhIFxUd\nfM4ARpnZ2ma2OlHLudvd5wLbEDWhrIfScoCtiWa6rOmZdBERqRFFB597ib6eV4EFwEhaai4Dgfll\n678ODMqZLiIiNaKwAQdm1ge4FbiZCDpvAQcB95jZVkATMLhssyHAM+nvJmCTVtIXVCO/gwb1p6Fh\nYDXeuipqKa+dUc/lq+eygconrStytNtQYDPgbHdflJZdbGanAzsAM4BhZdtsC1yT/p4B7FOWPiwt\n73YLFixm9uymarx1t2toGFgzee2Mei5fPZcNVL5aV83AWljwcfc5ZubAWDM7jpaazwDgYWAOcLOZ\nTQSmAuOAfsD16S0mA6eb2dHAOcAIoga1a1FlEBGR7lF0n89IYsTbi8BrxHDpL7j7C+4+FTgcuAiY\nB+wH7OHuCwHcfT6wJ7B/Sr8AGOPuDxRcBhER6aJCbzJ1dwf2bid9EjCpnfRpgO7rERGpcXq8joiI\nFE7BR0RECqfgIyIihVPwERGRwin4iIhI4RR8RESkcAo+IiJSOAUfEREpnIKPiIgUTsFHREQKp+Aj\nIiKFU/AREZHCKfiIiEjhFHxERKRwCj4iIlK4wubzMbNHgY3L9r06MMzdHzKz0cCJwHrAI8BYd5+e\n2X574FxgK2AWMN7drygq/yIi0n0Kq/m4+1buPqj0D/gF8FgKPMOB84AxwJrAdcAUMxsAYGaDgCnA\n1cAQYgbUC8xME8uJiNSgHml2M7O+wDeIqbABDgGudffb3P0dd58AvAnsm9JHAYvc/cyUfiswGTi0\n6LyLiEjX9VSfz77AIOD36fU2wLSydR5KywG2Bh4sS5+eSRcRkRrSU8HnUOAqd29KrwcC88vWeZ0I\nUHnSRUSkhhQ24KDEzLYAdgGy/TVNwOCyVYcAz2TSN2klfUE18ggwaFB/GhoGVuvtu10t5bUz6rl8\n9Vw2UPmkdYUHH2JQwUPu/u/MshnAsLL1tgWuyaTvU5Y+LC2vigULFjN7dlPHK/YCDQ0DayavnVHP\n5avnsoHKV+uqGVgLDT5mtipwEPCjsqQLgZvNbCIwFRgH9AOuT+mTgdPN7GjgHGAEMBLYtYh8i4hI\n9yq6z2c/YDXgyuxCd58KHA5cBMxL6+3h7gtT+nxgT2D/lH4BMMbdHygu6yIi0l0Krfm4+1XAVW2k\nTQImtbPtNJbvJxIRkRqlx+uIiEjhFHxERKRwCj4iIlK4TgUfM1vdzPp1d2ZERGTlkCv4mNkpZvbx\n9PduwFxgrpl9ppqZExGR+pS35nMQ8GT6+wTgB8BY4GfVyJSIiNS3vMFnkLsvMLP3Eg/zPN/dJwJb\nVi9rIiJSr/Le5zPHzD5ATOR2v7u/a2ZrVDFfIiJSx/IGn1/RMuXB19L/I4DHuz1HIiJS93I1u7n7\n2cBHga3cvfS8teeJh4SKiIhUpMOaT3oY6GvAuu7+Zmm5uz9VzYyJiEj96rDm4+7vEMFn1epnR0RE\nVgZ5R7v9BLjAzDaoZmZERGTlkHfAwaVAX+DLZrYUaC4luLuedCAiIhXJG3w0aZuIiHSbXMHH3e/s\nrh2a2a7AT4l7ht4A/uTu30lpo4ETgfWAR4Cx7j49s+32wLlp21nAeHe/orvyJiIixcj9YFEzO8DM\nbjGzh9PrEWa2XyU7M7NPA1cDZwBrAhsSs5diZsOB84jh22sC1wFTzGxASh8ETEnbDwEOI/qhNMGc\niEiNyVXzMbOjiGe5nUvUTABmE0Hkugr2dyrxaJ7JmWUPpf8PAa5199vS6wlmNhbYF7gcGAUscvcz\nU/qtZjYZOBS4v4I8iIhID8tb8zkM2MPdf0HLYIOnqODZbmbWH/g4sKqZTTOz2WZ2u5ltl1bZhpan\nKJQ8lJYDbA08WJY+PZMuIiI1Im/wGZq5qbQUfPpk/s5jzbS/LwOjgfWBW4CbzGwwMBCYX7bN68Cg\n9HdH6SIiUiPyjnZ73Mz2dve/ZJZ9FphRwb6a0v+XuPtj6e+fm9n3gU+l9MFl2wwBnslsv0kr6Qsq\nyENugwb1p6FhYDXeuipqKa+dUc/lq+eygconrcsbfI4naih/AlYzs3OIGszeeXeUpmR4oY3kpUQg\nG1a2fFvgmvT3DGCfsvRhVBYAc1uwYDGzZzd1vGIv0NAwsGby2hn1XL56LhuofLWumoE171Dru81s\nB6Lv5w6i+ezTmRpMXucBR5rZH4k+o6OBN4F7gYXAzWY2EZgKjAP6AaUHmU4GTjezo4FziKdqj0T3\nIImI1Jy8NR/c/XHgiK7szN3PTEOnbwdWIwYQ7OHuTcBUMzucGHpdus9nD3dfmLadb2Z7EgHsZOBl\nYIy7P9CVPImISPHyDrUeBTzm7k+a2ebAJcAS4FB3f7aSHbr7eGB8G2mTgEntbDsN0H09IiI1Lu9o\nt1NpGTBwBvAS8DTR/CUiIlKRvM1u67r7f82sL9HHsjHwFvDfquVMRETqVt6az1tmNgT4JPC0uy8g\nmt30RGsREalY3prPDcBtwADSs9iIJw68VI1MiYhIfcsbfL4DHAS8TcuAgMHE06lFREQqkvc+n7eB\nC8uW3VGVHImISN1rN/iY2e9aWfwO8BxwtbvPrEquRESkrnU04GDVVv4NBr4IPKq5dEREpDParfm4\n+8FtpZnZGOL+n126O1MiIlLfcs9k2orLgI90Uz5ERGQl0pXgAzGnj4iISEW6Eny+CjzaXRkREZGV\nR2dGu61KTJ89DNi9GpkSEZH61pnRbu8ANwNbu/s91c2eiIjUo06PdhMREems3JPJdZWZXQp8jZi5\ntA/QDBzr7hdk1hkNnEjLZHJj3X16Jn174FxgK2AWMN7dryiqDCIi0j0KCz7JZe5+aGsJZjacmKV0\nH+AuYhrtKWa2pbsvNLNBwBRiPqHhwI7AZDN7xt3vLyb7IiLSHbo61Lo7HQJc6+63ufs77j6BqCXt\nm9JHAYvc/cyUfiswGWg1mImISO/VZvAxs30zf6/aTfsbZWavmdmTZnaGmb03k7YNMK1s/YfScogp\nHB4sS5+eSRcRkRrRXs1nYubvOd2wr7OBD7j72kRtZkeWf1L2QGB+2TavA4NypouISI1or89nvpnt\nTnT8r2Jm69PKEw3cfVaeHbn7g5m/nzCzccCdZnaQu78DNBEPLc0aAjyT/m4CNmklfUGe/Vdq0KD+\nNDQMrMZbV0Ut5bUz6rl89Vw2UPmkde0Fnx8D1wJrpNf/KUsvjVjr24X9N9MS0GYQN65mbQtck0nf\npyx9WFre7RYsWMzs2U3VeOtu19AwsGby2hn1XL56LhuofLWumoG1zeDj7hPNbBKwPvAk8OGu7MjM\nvgT81d3nm9n7gTOBG9JEdRBNcDeb2URgKjHarR9wfUqfDJxuZkcD5wAjgJHArl3Jl4iIFK+jm0yX\nAP8xs13d/cUu7uvbwLlmthrwKnAdcFJmX1PN7HDgIlru89nD3Rem9PlmticxHPtk4GVgjLs/0MV8\niYhIwfJOo32fmX0M+AawEfAScIm7/yvvjtx9pxzrTAImtZM+DdAEdiIiNS7XfT5mNpK48XMwMdx5\nEDFYYN92NxQREWlF3icc/AQY5e5TSgvMbA/gNKIvRkREJLe8TzjYFPhr2bK/seLQZxERkQ7lDT4v\nsuKosl2Amd2bHRERWRnkbXb7KXCDmV0DPE/UhEYBB1UpXyIiUsdy1Xzc/VqiprMY2B54A9jN3a9p\nd0MREZFW5J5Swd3vBe6tYl5ERGQl0ZumVBARkZWEgo+IiBROwUdERArXYZ9PmkhuX+IhoG9VP0si\nIlLvOqz5pLl2LlLgERGR7pK32e3fZrZ1VXMiIiIrjbxDre8A/mxmvyOedrC0lODuV1YjYyIiUr/y\nBp9vEAHnkLLlzYCCj4iIVCTvfD6bdedOzawPMVvpDsCG7j4rLR8NnEjLZHJj3X16ZrvtgXOBrYBZ\nwHh3v6I78yYiItVX0VBrM3ufme3QDfs9ClhI1JxK7z2cmKV0DLAmMdPpFDMbkNIHAVOAq4EhwGHA\nBWamyeVERGpMrpqPma1DNK/tTDzfbYCZfQnY0d0Pr2SHZtZITKk9Cngok3QIcK2735ZeTzCzscQw\n78vT+ovqKigAAAAddUlEQVTc/cyUfquZTQYOBe6vJA8iItKz8tZ8ziaeZt0AvJOW3Q7sVsnOUnPb\nxcDRwPyy5G2AaWXLHkrLAbYmZlHNmp5JFxGRGpE3+OwEHOHuc0hNZe4+G1inwv2NA2a5+43pdTMt\nTW8DWTEgvU5M2Z0nXUREakTe0W5vla9rZkOBuXl3ZGZbEH0926VFfcr+bwIGl202BHgmk14+c+oQ\nYEHePFRi0KD+NDQMrMZbV0Ut5bUz6rl89Vw2UPmkdXmDz9+Bs8zsO5llJwE3VbCv4cDawKOp+W0V\nIvA8bGY/JprYhpVtsy1QmjNoBrBPWfqwtLzbLViwmNmzm6rx1t2uoWFgzeS1M+q5fPVcNlD5al01\nA2ve4HMscD0wD1jdzF6n9WDQnquAWzKvNwL+SfQbOTG0+mYzm0gMwx4H9Ev7BZgMnG5mRwPnACOA\nkaw4vbeIiPRyee/zmQuMSPfZbEo85eDf7t7c7obLv8ebxL05wLIHljYD/3P3xcBUMzscuIiW+3z2\ncPeFafv5ZrYnMRz7ZOBlYIy7P5A3DyIi0jvknskUwN3/bWYvuPtrXd2xu78I9C1bNgmY1M420wDd\n1yMiUuPy3ufTH/glcCCwmpm9BfweONrdF1UxfyIiUofyDrUuPdLmc0Aj8Hngw8BvqpQvERGpY3mb\n3T4HfDDd2wPwrJk9DDxRnWyJiEg9y1vzWQi8UbbsDeLeGxERkYrkrfmcCFxiZscCM4mbPX8OnFCt\njImISP1qM/iY2Ttknjqd1h2Ved0H2I946KeIiEhu7dV8dPOmiIhURZvBx93vLDIjIiKy8sh9k6mZ\nfRLYnni69DLufmp3Z0pEROpb3ptMTwG+TzzPbXEmqRlQ8BERkYrkrfmMAT7u7g9XMzMiIrJyyHuf\nzxvA49XMiIiIrDzyBp9fAD+uZkZERGTlkbfZ7WrgdjMbB7yaTXD3xm7PlYiI1LW8wecq4D/Ar1h+\nwEFF0sCFrwJrEU15dxFPxn4ppY8mnqZQms9nrLtPz2y/PS0POZ0FjHf3KzqbHxER6Rl5g89HgbXT\nhHBd8XvgdHdvMrPVgZ8BfwT+z8yGExPF7UMEpXHAFDPb0t0XmtkgYApwBjEl947AZDN7xt3v72K+\nRESkQHmDzxPAmsTsoZ3m7k9lXvYlhmqXmu0OAa5199vS6wlmNhbYl3iEzyhgkbufmdJvNbPJwKGA\ngo+ISA3JG3wuA641szOBV7IJ7n5vJTs0s68A5wODgHeA76WkbYBLy1Z/KC2/HNgaeLAsfTpwQCX7\nFxGRnpc3+JyT/r+mbHkzZVNhd8Td/wD8wczWAb4JPJqSBgLzy1Z/nQhSedJFRKRG5Ao+7p53SHZu\n7v6qmV0EPGdmGxNzAw0uW20I8Ez6u4mYyqE8fUF35w1g0KD+NDQM7HjFXqKW8toZ9Vy+ei4bqHzS\nutzPdquSVYH+wPrEo3uGlaVvS0ttawYxGCFrWFre7RYsWMzs2bUxV15Dw8CayWtn1HP56rlsoPLV\numoG1rzPdruF5ef2WcbdP5PzPfoAhwN/cvfZZrYh0Zz3PPAkcCFws5lNBKYSo936Adent5gMnG5m\nR6ftRgAj0dQPIiI1J29z2j1EQCj9ex7YjrgXpxJ7Ao+YWRPwT2J67t3cfam7TyWC00XAPGKiuj3c\nfSGAu89P2++f0i8Axrj7AxXmQUREeljePp+TypeZ2eXAEXl35O7NwF4drDMJmNRO+jTgE3n32VnN\nS5fy8sv/5dlnn65ou0033Zy+fSsafyEislLqSp/PVOCm7spIb7J4wf+4+p4F3PTIffm3mf8qvz7m\n82yxxfurmDMRkfrQqeBjZqsSN4W+1r3Z6T36D16HAWtu0NPZEBGpS3kHHLzD8gMO+hL9NQdXI1Mi\nIlLf8tZ8ykeUNQFPlQYDiIiIVCLvgIM7q50RERFZebQbfNIUB+1y9993X3ZERGRl0FHN54Q2ljcD\nDcRz1RR8RESkIu0GH3dfYdywmQ0lJnwbA0ysUr5ERKSO5R5qnYZXfxc4HvgX8Al3f7haGRMRkfqV\nd6j1l4GfE8Orv+ruf61qrkREpK51NOBgOHAmsCHR1HaZuy8tImMiIlK/Oqr53EU8xeBCYD3gh2a2\n3Arufmp1siYiIvUqT/BpBj7VRnozoOAjIiIV6Wi026cLyoeIiKxEun16bBERkY4UNo22mZ0G7A1s\nRDwbbgrwA3efl1lnNDGwYT1iorqx7j49k749cC6wFTALGO/uVxRVBhER6R5F1nzeBb4GDAW2IUbQ\nXVZKTCPrziNuXl0TuA6YYmYDUvogImBdDQwBDgMuMLOqTy4nIiLdq7Caj7v/OPNyjpn9Grgqs+wQ\n4Fp3vy29nmBmY4F9gcuBUcAidz8zpd9qZpOBQ4H7q5t7ERHpTj3Z57MrMCPzehtgWtk6D6XlAFsD\nD5alT8+ki4hIjSis5pNlZqOIGsuIzOKBwPyyVV8nHl6aJ11ERGpE4cHHzL4InA98zt2zNZ8mYHDZ\n6kOAZzLpm7SSvqAa+eyMoUMH0NAwsEf23VP7LUo9l6+eywYqn7Su0OBjZgcDE4C93f2+suQZwLCy\nZdsC12TS9ylLH8byTXc9au7chcye3VT4fhsaBvbIfotSz+Wr57KBylfrqhlYixxqfSQxjHp3dy/v\n24F4hM/NZjYRmAqMA/oB16f0ycDpZnY0cA7RZDeSFaf4FhGRXq7IAQe/Ivpt7jCzBWbWZGbLmszc\nfSpwOHARMA/YD9jD3Rem9PnAnsD+Kf0CYIy7P1BgGUREpBsUOdS6w0Dn7pOASe2kTwN0X4+ISI3T\n43VERKRwCj4iIlI4BR8RESmcgo+IiBROwUdERAqn4CMiIoVT8BERkcIp+IiISOF65KnW9ah56VJm\nznyx4u023XRz+vbtW4UciYj0Xgo+3eSNptmcddVr9B/8cu5tFs9/lV8f83m22OL9VcyZiEjvo+DT\njfoPXocBa27Q09kQEen11OcjIiKFU/AREZHCKfiIiEjhFHxERKRwRU+j/SVgLLANsIa79ytLH03M\ndroe8Agw1t2nZ9K3B84FtgJmAePd/YqCsi8iIt2k6JrPXCJ4jCtPMLPhwHnAGGBN4DpgipkNSOmD\ngCnA1cAQ4DDgAjPT5HIiIjWm0JqPu98CYGY7tpJ8CHCtu9+WXk8ws7HAvsDlwChgkbufmdJvNbPJ\nwKHA/dXNuYiIdKfe1OezDTCtbNlDaTnA1sCDZenTM+kiIlIjelPwGQjML1v2OjAoZ7qIiNSI3vSE\ngyZgcNmyIcAzmfRNWklfUOV8VdXQoQNoaBhY0TZLlizh2WefXfZ63rx8j/TZYostavY5cpV+RrWk\nnssGKp+0rjcFnxnAsLJl2wLXZNL3KUsflpbXrLlzFzJ7dlNF2zz77NN8d8KN9B+8Tu5tavk5cg0N\nAyv+jGpFPZcNVL5aV83AWvRQ61WAVYHV0uvVANz9LeBC4GYzmwhMJUbE9QOuT5tPBk43s6OBc4AR\nwEhg1yLL0FvoOXIiUsuKrvkcCFwKNKfXbwDNZraZu081s8OBi2i5z2cPd18I4O7zzWxPYjj2ycDL\nwBh3f6DgMnSbzk7D0JltRER6k6KHWk8EJraTPgmY1E76NKBu7uvpzDQMAHP+8wRrbfjBirbRfEMi\n0pv0pj6flVJnms8Wz/9fxfvRfEMi0pso+KxE1E8kIr1Fb7rPR0REVhIKPiIiUjgFHxERKZyCj4iI\nFE4DDqRNGp4tItWi4CNt0vBsEakWBR9pl4Zni0g1qM9HREQKp+AjIiKFU7ObdCsNUhCRPBR8pFtp\nkIKI5KHgI91OgxREpCPq8xERkcLVXM0nzYZ6OnAQMSPq34Fvu/ucHs2YdFpH/UTz5g1g7tyFKyxX\nPxEsWbKEF154ruLt9NlJT6u54AMcB3wO+Bgwl5gZ9XJgz57MlHReUf1EnT1RQ+89Wb/wwnN8d8KN\n9B+8Tu5t1McmvUEtBp9vAePd/UUAMzsWeMbMNnL3l3o2a9JZlfYTdWZU3cyZL3LWVTMqOlEDLHr9\nFb7/5W3ZeONNcm+zZMkSoA99+7bfsp2t1eXdJmvmzBcL+eyg9wZgqU01FXzMbDCwMTC9tMzdnzOz\nBcA2gILPSqIztaXS9OOdmTk2glZl+1pj4FoVBbrOblPplOqd+ew6E4ABhg7dpqL1obimRDVZ9qya\nCj7AQKAZmF+2/HVgUFsbrdL0FEvfem/unTTPn8XifpVdHb/RNBfoU/VtitxXb99mjYFrVbQNRJNT\npTq7r6JUWqbOlOfNhfM45cJbWH3A0Aq2mcuE7+/D4MENFe1r5swXO7WvH39rt4qCY3fsp63+yFrT\nE02wfZqbmwvfaWelms884KPu/nBm+evAAe7+lx7LnIiI5FZTQ63dfT4wExhWWmZmWxA1oofb2k5E\nRHqXmqr5AJjZ8cCBwB5ELehiYA1336tHMyYiIrnVWp8PwGnAEOBfQD/iPp8DezRHIiJSkZqr+YiI\nSO2rqT4fERGpDwo+IiJSOAUfEREpXC0OOMitVh9CamanAXsDGwFNwBTgB+4+L7POaOBEYD3gEWCs\nu0/PpG8PnAtsBcwiHkl0RWGFyMnM+gBTgR2ADd19Vlpe8+Uzs12BnxJ5fAP4k7t/J6XVdPnMbF3g\nbGAnoC/wIHBU6f67WiqfmX0JGEs8JWUNd+9Xlt6lsphZA/BbYFfiOLjU3X9Y1UItn/82y2dmBwLf\nBj4IvEsM5PqBuz+aWacq5av3mk/2IaQbErfRX96jOcrnXeBrwFDigNkQuKyUaGbDgfOAMcCawHXA\nFDMbkNIHEQHramJk4GHABWb2ieKKkNtRwELiyRVAfZTPzD5N5O8MogwbAheltJovH3A+kbctgXWB\nacBfoCbLN5c4uY4rT+imslwJLAXeB3wC2NfMjqlaaVbUZvmAAURg3SD9exD4u5mtDlUuX3Nzc93+\na2xsfKGxsfHrmdebNzY2Lm1sbNyop/NWYTl2b2xsfD3z+rLGxsaJrZT1wPT3wY2Njc+Xpf++sbHx\n4p4uS1meGhsbG59ubGzcOn0v76uX8jU2Nt7b2Nh4ahtp9VC+GY2Njd8q+y6XNDY2Dq3V8jU2Nu7Y\n2Nj4dnd+V42NjZulY3vTTPo3Ghsbn+0N5WtlndVSfj9a7fLVbc2nrYeQAqWHkNaSXYEZmdfbEFea\nWQ/RUq6tiSuYrOn0onKn5raLgaNZ8Vl9NV0+M+sPfBxY1cymmdlsM7vdzLZLq9R0+ZIzgFFmtna6\nSh4D3O3uc6mP8pV0tSxbA6+7+wtl6ZuWak+9zK7AIuDp9Lpq5avb4EMnH0La25jZKOBQ4MjM4oG0\nX66O0nuDccAsd78xvW6mpemt1su3JvHb+jIwGlgfuAW4KV0U1Xr5AO4l+npeJS7oRhLHKdRH+Uq6\nWpa20qGXldfMGoFLiL67RWlx1cpXz8GniejjGVy2fAjxY+n1zOyLREfe59w9W/Npov1ydZTeo9Lz\n+I4CjkiL+pT9X9PlI/IHcIm7P+bu77r7z4FVgU9R4+VLtdZbASdOPv2BU4F7zGwdarx8ZbpalrbS\nS2m9gpl9CLgdOMPdL8wkVa18dRt8av0hpGZ2MNGpu7e731WWPINMuZJtieaAUvpHy9KHsXzTXU8a\nDqwNPGpms4lmjT7Aw2b2baIcNVs+d18AvNBG8lJq//sbCmwGnO3ui1JwvZg4n+xA7Zcvq6tlmQEM\nNrNNM+nbAS+4e68IPmY2DLgDONXdzypLrlr56nqoNfA74Adm9g/iIaSnA39195k9mqsOmNmRxAiU\n3d29vL0Z4ELgZjObSAxTHkc85+76lD4ZON3MjgbOAUYQzSK7VjvvOV1FNEOVbAT8E9iNuJp+hNou\nH8QIqSPN7I/AU0Tf1ptEc9VCarh87j7HzBwYa2bHAW8RtzMMIC7s5lBD5Uu3ZKxK3I6Bma0G4O5v\n0cXfmru/YGa3AmeY2TeBBuBY4IJiStd++czs/4A/A8ekC4hyVStfXT/bLX3opwEH0/IQ0jGpU7TX\nMrOlwDvEjxqiVtDs7oMy6xwAnETLvQffdveHMunbESfArYCXgRPc/Q/FlKAyZrYJ8BywUeY+n5ov\nn5mNJzriVyM6bb+XuQ+mpstnZgacRQytfQ/wDHBSaU6tWiqfmR0EXEpLn2Of9Pdm7j6zq2Uxs7WJ\n5vPdiAuQi939uKoXrGX/bZaPuIVjBLCYlmbvZmAPd5+atq9K+eo6+IiISO9Ut30+IiLSeyn4iIhI\n4RR8RESkcAo+IiJSOAUfEREpnIKPiIgUTsFHREQKp+CzEjOzg8zs6Y7XlEqZWVNvmH/HzKaY2fd7\nOh+VMLOn0wRuvYKZXWpmv+vpfNSben+8Tq+SHvOzA/HkgqXEY0j+CfyqjcfoFKFLdxmb2fPAj9z9\nym7KT1Wlu70vIR4bD/HYpeuJx4u83V37cfeBOfPzE2C4u+/W1X2mJ2MMd/d7M/nYs6vv25ukp2E8\nT3x/zcSd+fcAR7v7iz2ZN6mMaj7FagZOdvfB7r4mMQXxC8A/zWyfHs3ZyuVZdx+UHlf0WWB/oLBp\njVvR7gWAma1aVEZqRDPQmL6/jwDrABM780Zm1jc9pVsKpppPD3L3l4ATzGx94qF9N8CyB4seRkxr\nO5eYpvZH7t5sZqcBH3D3kaX3MbOdiQcArkc8R+x3wM7E9/sS8SyqqW3lw8yOJR6YuAowiZjDfUlK\n2wj4BfEk6qXEVMlHufsiM7uRmLDvIjO7gHho5nHAXcAQd19iZt8gpo/e2d3/YWbrAv8F1nf32W28\n/9HuvjDtfygxcdlnUtnuAI5091dT+vOpvLsQzxl7nnh+3z9zfgdPmNndwPZln8lI4MfAFsS89T/L\n1u7SQxSPI57OfSPxXKx33f3glL6sFpKu1n+b8tdMPMfuK8TTkY8H+phZU0rbGtgx7fu3wHeJ+VE+\nYmY/I+YIWgd4BfiNu/867e+htP3f077/6O6HmtkdwC3ufmpab2vgl2nfc4lnfp2ajq1SrWJ0KtuG\nRM38IHf/X2ufn5l9Ka27GfHA1D8Tz7B7I6W3+/2Y2XuI7/drwBLgV21/W8vpA5COoauIKR0ws68T\nv5X3Z/J4KfBO+jxKZTyEeNjr5sAmZjaXeCDmaGI66P8Rv4Pr0tusnprevpjK+VN3/116/w2IY3w7\n4gGejwDj3H16Sv8o8fv+SCrjE8Be7j7fzPoCPyAezNoAPJa27amWkMKo5tM7/BHYIE3mBBEwdk9X\ndvsA3yB+LBA/5M+mk3jJN4FJ6Qd/DLAG8ZDOIcC+wH/a2femxFOlNwU+CXwuvUfp6be3A48CmwAf\nIgLi2QDu/nli2opvpprEZ939QeCN9F4QT799mpYnFu8KPJpOGm29/68z+bue+MF+KK3TRATjrIOB\n7xCTV91KBVfBZrYNcbJ/MrNsN+JpxkemGupBwG/MbHhKH0GcTL5JTC8whag9tVWDORV4kTi5rAV8\nHZjn7n9Kaf9w94HpM3whbbMpcTGxJfCxtOwx4FOpSe9bwM9TXnH3jxIn5N3S+5QmdsuWdRDxcN3b\ngHWBvYlj66iyVfcnLgY2IJ5UfXIb5YIIjF9x98HA/0vb/bhsnfa+n+OAPYnm6M1SuTduZ3/lZXof\nEZAfSIuykxK25yvAp4kpVl4DfgZ8FRiVfnc7Ek8jLxkF3JCOhyOJ42GjlLYKcC7xO1qPmCLkuhRY\nSGl/S7/HdYjPu9TEezLxm/sMcWxcAvzVYtLBuqbg0zv8hzhxrAXg7pNL0z6kSeQuJ64cS1OB30Wc\nEDGzNYkAU5oA6u30Ph80sz7u/kwHbeFLgO+7+9vu/jxxFfr1lPa5tM+TUvp84CfA18qaKsqbLe6g\nJdjsTJyMSn0auxAnoA7f38y2J+YO+Y67L3T3N4nmsZ3TSafkAnd/0t2biSvQLcysvT6Xzc1srpkt\nJp42fTfx1OKSI4Ffl/pO3P3fRI2w1Al+IPAnd7/T3Ze6+x+B+9vZ39ukQOLuze7+qLu/1s76pW1+\n6O5vpXLj7leWaiDu/g/gJtJxkdFeE9LewFvufqq7v+PuTxLTjBxStt54d5+Xap9XUlYrzHL3v7n7\nE+nv54g5qMrz1N73cyBwmrs/7zGFQZ7BEX2Ax1Jt5Z9EYD8ox3ZZ4919tsdcREuBw4nfwWOpLLPc\n/dHM+re7+00pbTIRdD+aXr/k7n9J39VbxHQoGwOl2tfbwMZmtom7L3H3B0o1Q2JCxWPc/cV0bFxK\nPDl6rwrLU3PU7NY7bEhcrc0BMLOvAN8jmgT6EtNB3JdZ/3fAKUSgOAB43Fse8X4G8b1OBNYzs78Q\nzQevtrHvV9MPpuSFlB+Iq9BSk0RJHyJgrUf8SFpzK3CQmV1HdOhfC5yfmtB2IaYZyPP+mwKrA/+L\nJ/gvS19M/LhnpWWvZLbPTv/b1mRWz7l7YwqgXyOm3RiS2XYz4NNmVqoR9CEu1EqT+m0A/KvsPdsL\n8N8nTkh/NrP+xOfxQ3df3M42L7v7u9kFqTn2EOL76UN8NpUM9NiwlXw+S1yxlzSz4ufZZiBPNa8T\ngA8Qx+l7iCarrPa+n+Xy5O6LzaytYzWbxw+5e1vHX0eas/s0swbgvUQNvS3l+1r2uZjZWkTT8aeJ\nWT1Lta8Gokb9deL7v8fM3gauAMYTteYBxHGRne7gPbT8BuuWgk/v8GXgv+7+lJltSNR0RgI3p36T\nCUR7csn1wNmp+ecbxNUmAOmK6gSiL2kd4kDP1mbKrWNmq5eurokTb6mZ7sV4S/9IO3lf2sqyW4mm\nhv2I/oalqV/lMKK55848729mLwIL3X1oO/vvtHQlPsnMdiWa0fbL5OtSX3FWx5L/Ek2AWRsTJ/LW\n9jOH6Lv5rsWMjzcS/Qvjaf3zo3y5mX2KCJI7ufv9adnVLF/T6ai56aVW8r1FWl6xNBBiMhFcL3H3\nt81sLNGXktd/iYuM0nu+lzhpd6StGl4TEUiy3seKQXfZ55uagBcRNZVWv8MOnEZcLH3M3V81swHE\nNNOlfqkXiSZazOzDxESKz7n7ZWa2ENh1ZejjKadmtx5kZhua2UlEc86RafEA4qB9LQWeHYimiWXS\nFfFlRMfxlkB2Yqe9zewDFhPpLSYmd1rSTjb6EjMVrm5mmxMnjstS2l+AfmZ2XPpBYWYbpM74kldo\naV4o5e954oQ2jpYZS28n+pLuyzQ5dPT+/wZmmNk5qdaEmTWkTu72VDp66SRgLzP7eHr9K+B7Zjbc\nzFYxs35mNsxiUi2Ii4MvmNmOKf1LRJ9Fq8xsf2uZZriJaIYpfSevEE0yHY1oGwS8C7yWmiT3AvYo\nW+dlyr6LMjcBq6XPe1WL6uSxRFNYSSWfXb/07/UUeD5E9O10JLuPy4FjzGxzM1uDuFDqKA/tpT9E\nXFDtmT6nfYnJ0jra/nxiNs4Pw7LjsL2LrqyBxG9tfjqOzyBzIWBmoy0GFUEEpXfTP4j+07PMbMu0\n7gAz+4yZrZdz3zVLwad4J5jZfDN7nagBbA580t2vB0jt8D8BbjSzecTJobWmlYuINuerfPm50rcg\nRhzNJ0ZVLSZG07TlBaKm8zzRfj4FmJDy8gbRZ/Mh4MmU51uAbTLbnwIcaGZzzOymzPJbiWahOzKv\nB5KZPruj9081k32Ik8U0M5tPjKjbMbOf1q72K7p3KQXLicDP0+tbiA79CURn9H+JZpX3pvS7iJrM\npcSIsT2JGkC2+TKbh22BOy1GtD1CBNUJKe1qIlC/kvqhymsmJX8Dfk80980mamnXla3zI+Cn6bso\n1YaX5cPdFxAd27sRTWM303IR01q+2+Xui4ja7AQzW0DUHq8oW62j7+fnqWz3EbWOF2i/CbPdPKZ+\np+8SfaBziPJek2P744E/Aden7+kO4reUJw8nEjX6OUTwu4flL/h2Jo7fJmIq7knuPimz7fXADen4\nd6JZuu7PzZrJtEalvoNXgV1KzTDSc8zsXuBGdz+tp/MiUgvU51ODUkf594DHFHh6hpmNAv4KvEP0\np21HWfOoiLRNwafGpJE5zxHNJl/s4eyszEYRTZ+rAM8AI929M53VIislNbuJiEjh6r5TS0REeh8F\nHxERKZyCj4iIFE7BR0RECqfgIyIihVPwERGRwv1/rS5s7S6FNnMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x14783be90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tmp.diff_acc_to_purchase.hist(bins=25)\n",
    "plt.title(\"\\nDays since Registration to make First Purchase\\n\", fontsize='small')\n",
    "plt.xlabel(\"Days between Registration and Purchase\", fontsize=13)\n",
    "plt.ylabel(\"Number of Users\", fontsize=13)\n",
    "plt.savefig('images/days_from_reg_to_purch2.png', bbox_inches='tight', pad_inches=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 772,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_sessions</th>\n",
       "      <th>total_artists_followed</th>\n",
       "      <th>total_artworks_favorited</th>\n",
       "      <th>total_artworks_shared</th>\n",
       "      <th>gender_male</th>\n",
       "      <th>gender_unknown</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</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>11680.0</th>\n",
       "      <td>224.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1198.0</th>\n",
       "      <td>462.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12879.0</th>\n",
       "      <td>207.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>98.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14105.0</th>\n",
       "      <td>266.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14184.0</th>\n",
       "      <td>821.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15149.0</th>\n",
       "      <td>538.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15608.0</th>\n",
       "      <td>653.0</td>\n",
       "      <td>136.0</td>\n",
       "      <td>203.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15993.0</th>\n",
       "      <td>471.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>190.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16963.0</th>\n",
       "      <td>533.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17100.0</th>\n",
       "      <td>1085.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17440.0</th>\n",
       "      <td>273.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1758.0</th>\n",
       "      <td>377.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17700.0</th>\n",
       "      <td>773.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>495.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18277.0</th>\n",
       "      <td>4451.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>994.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18327.0</th>\n",
       "      <td>239.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18366.0</th>\n",
       "      <td>690.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>97.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18973.0</th>\n",
       "      <td>288.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>333.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2002.0</th>\n",
       "      <td>269.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20780.0</th>\n",
       "      <td>2715.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2120.0</th>\n",
       "      <td>228.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2185.0</th>\n",
       "      <td>246.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>88.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2230.0</th>\n",
       "      <td>222.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23356.0</th>\n",
       "      <td>528.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23927.0</th>\n",
       "      <td>1861.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>98.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26269.0</th>\n",
       "      <td>594.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26436.0</th>\n",
       "      <td>456.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26710.0</th>\n",
       "      <td>300.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2698.0</th>\n",
       "      <td>298.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27057.0</th>\n",
       "      <td>483.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27802.0</th>\n",
       "      <td>1374.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>98.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70284.0</th>\n",
       "      <td>257.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72069.0</th>\n",
       "      <td>245.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72381.0</th>\n",
       "      <td>205.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73581.0</th>\n",
       "      <td>226.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73958.0</th>\n",
       "      <td>938.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74045.0</th>\n",
       "      <td>789.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74486.0</th>\n",
       "      <td>388.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74505.0</th>\n",
       "      <td>277.0</td>\n",
       "      <td>97.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74795.0</th>\n",
       "      <td>285.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75033.0</th>\n",
       "      <td>336.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75711.0</th>\n",
       "      <td>214.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76093.0</th>\n",
       "      <td>226.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76451.0</th>\n",
       "      <td>311.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77445.0</th>\n",
       "      <td>251.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>93.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77579.0</th>\n",
       "      <td>896.0</td>\n",
       "      <td>93.0</td>\n",
       "      <td>98.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77906.0</th>\n",
       "      <td>236.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79021.0</th>\n",
       "      <td>330.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79381.0</th>\n",
       "      <td>227.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80573.0</th>\n",
       "      <td>363.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81194.0</th>\n",
       "      <td>226.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81491.0</th>\n",
       "      <td>237.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82156.0</th>\n",
       "      <td>291.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82426.0</th>\n",
       "      <td>461.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84659.0</th>\n",
       "      <td>854.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85163.0</th>\n",
       "      <td>1564.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85239.0</th>\n",
       "      <td>234.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88024.0</th>\n",
       "      <td>483.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89352.0</th>\n",
       "      <td>245.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>910.0</th>\n",
       "      <td>3122.0</td>\n",
       "      <td>98.0</td>\n",
       "      <td>511.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9407.0</th>\n",
       "      <td>498.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>706.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>304 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         num_sessions  total_artists_followed  total_artworks_favorited  \\\n",
       "id                                                                        \n",
       "11680.0         224.0                     9.0                      94.0   \n",
       "1198.0          462.0                     7.0                      99.0   \n",
       "12879.0         207.0                    51.0                      98.0   \n",
       "14105.0         266.0                     8.0                      96.0   \n",
       "14184.0         821.0                     1.0                       9.0   \n",
       "15149.0         538.0                    16.0                      34.0   \n",
       "15608.0         653.0                   136.0                     203.0   \n",
       "15993.0         471.0                     8.0                     190.0   \n",
       "16963.0         533.0                     0.0                      99.0   \n",
       "17100.0        1085.0                    99.0                      40.0   \n",
       "17440.0         273.0                     9.0                       9.0   \n",
       "1758.0          377.0                     3.0                      94.0   \n",
       "17700.0         773.0                    99.0                     495.0   \n",
       "18277.0        4451.0                    99.0                     994.0   \n",
       "18327.0         239.0                     3.0                      30.0   \n",
       "18366.0         690.0                     7.0                      97.0   \n",
       "18973.0         288.0                    31.0                     333.0   \n",
       "2002.0          269.0                     9.0                      99.0   \n",
       "20780.0        2715.0                     8.0                       8.0   \n",
       "2120.0          228.0                     1.0                      67.0   \n",
       "2185.0          246.0                     5.0                      88.0   \n",
       "2230.0          222.0                     7.0                       8.0   \n",
       "23356.0         528.0                    26.0                      99.0   \n",
       "23927.0        1861.0                    96.0                      98.0   \n",
       "26269.0         594.0                     9.0                       9.0   \n",
       "26436.0         456.0                     9.0                      40.0   \n",
       "26710.0         300.0                     5.0                      46.0   \n",
       "2698.0          298.0                     3.0                      99.0   \n",
       "27057.0         483.0                     1.0                      68.0   \n",
       "27802.0        1374.0                    65.0                      98.0   \n",
       "...               ...                     ...                       ...   \n",
       "70284.0         257.0                     9.0                       0.0   \n",
       "72069.0         245.0                     2.0                       4.0   \n",
       "72381.0         205.0                     9.0                      80.0   \n",
       "73581.0         226.0                     2.0                       3.0   \n",
       "73958.0         938.0                     5.0                       9.0   \n",
       "74045.0         789.0                     9.0                       7.0   \n",
       "74486.0         388.0                     9.0                       9.0   \n",
       "74505.0         277.0                    97.0                      94.0   \n",
       "74795.0         285.0                     3.0                       8.0   \n",
       "75033.0         336.0                     1.0                       9.0   \n",
       "75711.0         214.0                     2.0                       9.0   \n",
       "76093.0         226.0                     1.0                       6.0   \n",
       "76451.0         311.0                     1.0                       5.0   \n",
       "77445.0         251.0                     8.0                      93.0   \n",
       "77579.0         896.0                    93.0                      98.0   \n",
       "77906.0         236.0                     3.0                       9.0   \n",
       "79021.0         330.0                     1.0                       9.0   \n",
       "79381.0         227.0                     0.0                       8.0   \n",
       "80573.0         363.0                     4.0                       8.0   \n",
       "81194.0         226.0                     8.0                       9.0   \n",
       "81491.0         237.0                     9.0                       9.0   \n",
       "82156.0         291.0                     9.0                      96.0   \n",
       "82426.0         461.0                     9.0                      96.0   \n",
       "84659.0         854.0                     7.0                       7.0   \n",
       "85163.0        1564.0                     1.0                       9.0   \n",
       "85239.0         234.0                     9.0                       9.0   \n",
       "88024.0         483.0                     6.0                       0.0   \n",
       "89352.0         245.0                     4.0                      99.0   \n",
       "910.0          3122.0                    98.0                     511.0   \n",
       "9407.0          498.0                    99.0                     706.0   \n",
       "\n",
       "         total_artworks_shared  gender_male  gender_unknown  \n",
       "id                                                           \n",
       "11680.0                    3.0          0.0             1.0  \n",
       "1198.0                     2.0          0.0             1.0  \n",
       "12879.0                    0.0          0.0             1.0  \n",
       "14105.0                    0.0          1.0             0.0  \n",
       "14184.0                    0.0          0.0             1.0  \n",
       "15149.0                    9.0          0.0             1.0  \n",
       "15608.0                    0.0          0.0             1.0  \n",
       "15993.0                    9.0          1.0             0.0  \n",
       "16963.0                    0.0          0.0             1.0  \n",
       "17100.0                    1.0          0.0             1.0  \n",
       "17440.0                    0.0          0.0             1.0  \n",
       "1758.0                     2.0          0.0             1.0  \n",
       "17700.0                    9.0          0.0             0.0  \n",
       "18277.0                    0.0          0.0             1.0  \n",
       "18327.0                    1.0          1.0             0.0  \n",
       "18366.0                    2.0          1.0             0.0  \n",
       "18973.0                    0.0          0.0             1.0  \n",
       "2002.0                     0.0          1.0             0.0  \n",
       "20780.0                    2.0          0.0             1.0  \n",
       "2120.0                     0.0          0.0             0.0  \n",
       "2185.0                     0.0          0.0             0.0  \n",
       "2230.0                     0.0          0.0             1.0  \n",
       "23356.0                    0.0          0.0             1.0  \n",
       "23927.0                    8.0          0.0             0.0  \n",
       "26269.0                    0.0          0.0             1.0  \n",
       "26436.0                    0.0          0.0             1.0  \n",
       "26710.0                    2.0          0.0             0.0  \n",
       "2698.0                     7.0          0.0             1.0  \n",
       "27057.0                    0.0          0.0             1.0  \n",
       "27802.0                    9.0          1.0             0.0  \n",
       "...                        ...          ...             ...  \n",
       "70284.0                    1.0          1.0             0.0  \n",
       "72069.0                    0.0          0.0             1.0  \n",
       "72381.0                    0.0          0.0             1.0  \n",
       "73581.0                    0.0          0.0             1.0  \n",
       "73958.0                    0.0          0.0             1.0  \n",
       "74045.0                    0.0          0.0             1.0  \n",
       "74486.0                    6.0          0.0             0.0  \n",
       "74505.0                    1.0          0.0             1.0  \n",
       "74795.0                    0.0          1.0             0.0  \n",
       "75033.0                    0.0          0.0             1.0  \n",
       "75711.0                    0.0          1.0             0.0  \n",
       "76093.0                    0.0          0.0             1.0  \n",
       "76451.0                    2.0          1.0             0.0  \n",
       "77445.0                    2.0          0.0             1.0  \n",
       "77579.0                    9.0          0.0             0.0  \n",
       "77906.0                    9.0          1.0             0.0  \n",
       "79021.0                    1.0          1.0             0.0  \n",
       "79381.0                    0.0          0.0             1.0  \n",
       "80573.0                    0.0          1.0             0.0  \n",
       "81194.0                    5.0          0.0             1.0  \n",
       "81491.0                    0.0          0.0             1.0  \n",
       "82156.0                    0.0          1.0             0.0  \n",
       "82426.0                    0.0          0.0             1.0  \n",
       "84659.0                    0.0          0.0             1.0  \n",
       "85163.0                    0.0          1.0             0.0  \n",
       "85239.0                    0.0          0.0             1.0  \n",
       "88024.0                    0.0          0.0             1.0  \n",
       "89352.0                    0.0          0.0             1.0  \n",
       "910.0                      0.0          0.0             1.0  \n",
       "9407.0                     0.0          0.0             1.0  \n",
       "\n",
       "[304 rows x 6 columns]"
      ]
     },
     "execution_count": 772,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.num_sessions > 200]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}