checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from pandas.io.json import json_normalize\n",
    "import numpy as np\n",
    "import sys, re, json, os, datetime\n",
    "from collections import Counter\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n",
      "CPU times: user 3min 51s, sys: 27.8 s, total: 4min 19s\n",
      "Wall time: 4min 26s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "def normalize_name(name):\n",
    "    name = name.lower()\n",
    "    name = re.sub(r\"[.,\\/#!$%\\^\\*;:{}=\\_`~()@-]\", ' ', name)\n",
    "    name = re.sub(r\"&\", 'and', name)\n",
    "    name = re.sub(r'\\s+', ' ', name).strip()\n",
    "    return name\n",
    "\n",
    "def make_social_df(path, num_slices, cutoff):\n",
    "    df_list = []\n",
    "    count = 0\n",
    "    filenames = os.listdir(path)\n",
    "    for filename in sorted(filenames):\n",
    "        if count < num_slices and filename.startswith(\"mpd.slice.\") and filename.endswith(\".json\"):\n",
    "            if count % 100 == 0:\n",
    "                print(count)\n",
    "            fullpath = os.sep.join((path, filename))\n",
    "            data = json.load(open(fullpath))\n",
    "            slice_df = json_normalize(data, 'playlists')\n",
    "            slice_df['normalized_name'] = slice_df['name'].apply(lambda x: normalize_name(x))\n",
    "            popular_df = slice_df[slice_df['num_followers'] >= cutoff]\n",
    "            df_list.append(popular_df)\n",
    "            count += 1\n",
    "    return pd.concat(df_list)\n",
    "\n",
    "path = '../mpd.v1/data'\n",
    "df = make_social_df(path, 1000, 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(29651, 13)\n"
     ]
    }
   ],
   "source": [
    "feature_cols = ['duration_ms', 'num_albums', 'num_artists', 'num_edits', 'num_followers', 'num_tracks', 'collaborative']\n",
    "df = df.replace({'collaborative': {'false':0, 'true':1}})\n",
    "df[feature_cols] = df[feature_cols].apply(pd.to_numeric, errors='coerce')\n",
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "const             3.795992\n",
       "duration_ms      18.742227\n",
       "num_albums       24.577429\n",
       "num_artists      12.561705\n",
       "num_edits         1.885403\n",
       "num_followers     1.001204\n",
       "num_tracks       23.130430\n",
       "collaborative     1.005172\n",
       "dtype: float64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
    "from statsmodels.tools.tools import add_constant\n",
    "\n",
    "df_w_constant = add_constant(df[feature_cols])\n",
    "\n",
    "pd.Series([variance_inflation_factor(df_w_constant.values, i) for i in range(df_w_constant.shape[1])], index=df_w_constant.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/scipy/linalg/basic.py:1226: RuntimeWarning: internal gelsd driver lwork query error, required iwork dimension not returned. This is likely the result of LAPACK bug 0038, fixed in LAPACK 3.2.2 (released July 21, 2010). Falling back to 'gelss' driver.\n",
      "  warnings.warn(mesg, RuntimeWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model accuracy: 0.0012024154423125877\n",
      "MSE: 557745.1426921646\n",
      "MAE: 71.10511250007556, 95% CI [64 77]\n",
      "R2: 0.0006246481968968931\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_predict, cross_val_score\n",
    "from sklearn import linear_model, metrics\n",
    "import numpy as np\n",
    "import scipy.stats as st\n",
    "import seaborn as sns\n",
    "\n",
    "data = df[feature_cols]\n",
    "\n",
    "X = data.drop(['num_followers'],axis=1)\n",
    "y = df['num_followers'] \n",
    "\n",
    "lm = linear_model.LinearRegression(normalize=True)\n",
    "\n",
    "model = lm.fit(X, y)\n",
    "model_score = model.score(X, y)\n",
    "print('Model accuracy:', model_score)\n",
    "\n",
    "y_pred = cross_val_predict(lm, X, y, cv=10)\n",
    "\n",
    "mae = cross_val_score(lm, X, y, scoring='neg_mean_absolute_error', cv=10)\n",
    "mae_ci = np.array((0,2))\n",
    "mae_ci[1], mae_ci[0] = np.abs(st.t.interval(0.95, len(mae)-1, loc=np.mean(mae), scale=st.sem(mae)))\n",
    "\n",
    "print('MSE:', metrics.mean_squared_error(y, y_pred))\n",
    "print('MAE: {0}, 95% CI {1}'.format(metrics.mean_absolute_error(y, y_pred), mae_ci))\n",
    "print('R2:', metrics.r2_score(y, y_pred))\n",
    "\n",
    "plt.scatter(y, y_pred)\n",
    "plt.xlabel('True counts')\n",
    "plt.ylabel('Predicted counts')\n",
    "plt.title('Predicted vs. True Counts \\n of Spotify Playlist Followers Regression Model')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/sklearn/cross_validation.py:41: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(29651, 13)\n"
     ]
    }
   ],
   "source": [
    "from imblearn.over_sampling import SMOTE\n",
    "from sklearn.linear_model import LogisticRegressionCV\n",
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn.metrics import precision_recall_fscore_support, confusion_matrix, roc_curve, auc\n",
    "\n",
    "print(df.shape)\n",
    "\n",
    "X = df[feature_cols].drop('collaborative', axis=1)\n",
    "y = df['collaborative']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=123)\n",
    "\n",
    "sm = SMOTE(random_state=123)\n",
    "X_train_smote, y_train_smote = sm.fit_sample(X_train, y_train)\n",
    "\n",
    "model = LogisticRegressionCV()\n",
    "fitted = model.fit(X_train_smote, y_train_smote)\n",
    "y_pred = fitted.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The model selected [1291.54966501] as C.\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegressionCV\n",
    "\n",
    "model = LogisticRegressionCV(Cs=10, random_state=0)\n",
    "fitted = model.fit(X_train_smote, y_train_smote)\n",
    "y_pred = fitted.predict(X_test)\n",
    "\n",
    "print('The model selected {0} as C.'.format(fitted.C_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision, recall, and F1 score, averaged and weighted by number of instances in each class:\n",
      "\n",
      "precision: 0.9486532662344496\n",
      "recall: 0.16928005395380205\n",
      "f1 score: 0.24317727195776712\n",
      "\n",
      "Precision, recall, and F1 score, per class [0 1]:\n",
      "precision: [0.98081535 0.03649205]\n",
      "recall: [0.14278234 0.92079208]\n",
      "f1 score: [0.24927625 0.07020192]\n"
     ]
    }
   ],
   "source": [
    "precision, recall, fscore, _ = precision_recall_fscore_support(y_test, y_pred, average='weighted')\n",
    "\n",
    "print('Precision, recall, and F1 score, averaged and weighted by number of instances in each class:\\n')\n",
    "print('precision: {}'.format(precision))\n",
    "print('recall: {}'.format(recall))\n",
    "print('f1 score: {}\\n'.format(fscore))\n",
    "\n",
    "precision, recall, fscore, _ = precision_recall_fscore_support(y_test, y_pred)\n",
    "\n",
    "print('Precision, recall, and F1 score, per class [0 1]:')\n",
    "print('precision: {}'.format(precision))\n",
    "print('recall: {}'.format(recall))\n",
    "print('f1 score: {}'.format(fscore))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x17c778198>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, y_pred)\n",
    "sns.heatmap(cm, annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x160f3fe80>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Compute fpr, tpr, thresholds and roc auc\n",
    "fpr, tpr, thresholds = roc_curve(y_test, y_pred)\n",
    "roc_auc = np.trapz(tpr,fpr)\n",
    "\n",
    "# Plot ROC curve\n",
    "plt.plot(fpr, tpr, label='ROC curve (area = %0.3f)' % roc_auc)\n",
    "plt.plot([0, 1], [0, 1], 'k--')  # random predictions curve\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.0])\n",
    "plt.xlabel('False Positive Rate or (1 - Specifity)')\n",
    "plt.ylabel('True Positive Rate or (Sensitivity)')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}