spotify-song-classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Intial setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Imports\n",
    "\n",
    "import spotipy\n",
    "from spotipy.oauth2 import SpotifyClientCredentials\n",
    "import spotipy.util as util\n",
    "import pandas as pd\n",
    "from pandas import read_csv\n",
    "import torch\n",
    "import numpy as np\n",
    "from sklearn import preprocessing\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn import decomposition\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>acousticness</th>\n",
       "      <th>artists</th>\n",
       "      <th>danceability</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>energy</th>\n",
       "      <th>explicit</th>\n",
       "      <th>id</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>key</th>\n",
       "      <th>liveness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>mode</th>\n",
       "      <th>name</th>\n",
       "      <th>popularity</th>\n",
       "      <th>release_date</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>tempo</th>\n",
       "      <th>valence</th>\n",
       "      <th>year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>27405</th>\n",
       "      <td>0.506</td>\n",
       "      <td>['The Jimmy Castor Bunch']</td>\n",
       "      <td>0.645</td>\n",
       "      <td>222401</td>\n",
       "      <td>0.8450</td>\n",
       "      <td>0</td>\n",
       "      <td>1zFg9JAeAPnGynPGXa5jhY</td>\n",
       "      <td>0.553000</td>\n",
       "      <td>7</td>\n",
       "      <td>0.1690</td>\n",
       "      <td>-10.960</td>\n",
       "      <td>1</td>\n",
       "      <td>It's Just Begun</td>\n",
       "      <td>32</td>\n",
       "      <td>1972-03-04</td>\n",
       "      <td>0.0600</td>\n",
       "      <td>117.543</td>\n",
       "      <td>0.8610</td>\n",
       "      <td>1972</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119814</th>\n",
       "      <td>0.881</td>\n",
       "      <td>['Hank Williams']</td>\n",
       "      <td>0.512</td>\n",
       "      <td>167933</td>\n",
       "      <td>0.3330</td>\n",
       "      <td>0</td>\n",
       "      <td>4tYNTVSL2alrVEFBCx4NRl</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0</td>\n",
       "      <td>0.2110</td>\n",
       "      <td>-11.685</td>\n",
       "      <td>1</td>\n",
       "      <td>Dear Brother - Single Version</td>\n",
       "      <td>17</td>\n",
       "      <td>1956-01-01</td>\n",
       "      <td>0.0301</td>\n",
       "      <td>110.689</td>\n",
       "      <td>0.3290</td>\n",
       "      <td>1956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>155720</th>\n",
       "      <td>0.682</td>\n",
       "      <td>['Vince Guaraldi']</td>\n",
       "      <td>0.609</td>\n",
       "      <td>228067</td>\n",
       "      <td>0.5230</td>\n",
       "      <td>0</td>\n",
       "      <td>1MSHhgOwXXcdP2EoR6tEKW</td>\n",
       "      <td>0.943000</td>\n",
       "      <td>7</td>\n",
       "      <td>0.1200</td>\n",
       "      <td>-13.537</td>\n",
       "      <td>0</td>\n",
       "      <td>Chora Tua Tristeza</td>\n",
       "      <td>19</td>\n",
       "      <td>1963</td>\n",
       "      <td>0.0312</td>\n",
       "      <td>104.643</td>\n",
       "      <td>0.9440</td>\n",
       "      <td>1963</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163305</th>\n",
       "      <td>0.575</td>\n",
       "      <td>['Los Hermanos Arriagada']</td>\n",
       "      <td>0.593</td>\n",
       "      <td>163579</td>\n",
       "      <td>0.3060</td>\n",
       "      <td>0</td>\n",
       "      <td>4gIFUzWudA3kQIDl3AhpOn</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4</td>\n",
       "      <td>0.0910</td>\n",
       "      <td>-15.163</td>\n",
       "      <td>0</td>\n",
       "      <td>Caminemos</td>\n",
       "      <td>9</td>\n",
       "      <td>1954-01-19</td>\n",
       "      <td>0.0299</td>\n",
       "      <td>101.856</td>\n",
       "      <td>0.5470</td>\n",
       "      <td>1954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>153978</th>\n",
       "      <td>0.977</td>\n",
       "      <td>['Aaron Copland', 'William Kapell']</td>\n",
       "      <td>0.534</td>\n",
       "      <td>458293</td>\n",
       "      <td>0.0648</td>\n",
       "      <td>0</td>\n",
       "      <td>2IfzKP18U3Hhzn9yIvzFPI</td>\n",
       "      <td>0.824000</td>\n",
       "      <td>6</td>\n",
       "      <td>0.0841</td>\n",
       "      <td>-18.247</td>\n",
       "      <td>1</td>\n",
       "      <td>Piano Sonata: Molto moderato</td>\n",
       "      <td>0</td>\n",
       "      <td>1945</td>\n",
       "      <td>0.0866</td>\n",
       "      <td>106.181</td>\n",
       "      <td>0.0448</td>\n",
       "      <td>1945</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        acousticness                              artists  danceability  \\\n",
       "27405          0.506           ['The Jimmy Castor Bunch']         0.645   \n",
       "119814         0.881                    ['Hank Williams']         0.512   \n",
       "155720         0.682                   ['Vince Guaraldi']         0.609   \n",
       "163305         0.575           ['Los Hermanos Arriagada']         0.593   \n",
       "153978         0.977  ['Aaron Copland', 'William Kapell']         0.534   \n",
       "\n",
       "        duration_ms  energy  explicit                      id  \\\n",
       "27405        222401  0.8450         0  1zFg9JAeAPnGynPGXa5jhY   \n",
       "119814       167933  0.3330         0  4tYNTVSL2alrVEFBCx4NRl   \n",
       "155720       228067  0.5230         0  1MSHhgOwXXcdP2EoR6tEKW   \n",
       "163305       163579  0.3060         0  4gIFUzWudA3kQIDl3AhpOn   \n",
       "153978       458293  0.0648         0  2IfzKP18U3Hhzn9yIvzFPI   \n",
       "\n",
       "        instrumentalness  key  liveness  loudness  mode  \\\n",
       "27405           0.553000    7    0.1690   -10.960     1   \n",
       "119814          0.000002    0    0.2110   -11.685     1   \n",
       "155720          0.943000    7    0.1200   -13.537     0   \n",
       "163305          0.000000    4    0.0910   -15.163     0   \n",
       "153978          0.824000    6    0.0841   -18.247     1   \n",
       "\n",
       "                                 name  popularity release_date  speechiness  \\\n",
       "27405                 It's Just Begun          32   1972-03-04       0.0600   \n",
       "119814  Dear Brother - Single Version          17   1956-01-01       0.0301   \n",
       "155720             Chora Tua Tristeza          19         1963       0.0312   \n",
       "163305                      Caminemos           9   1954-01-19       0.0299   \n",
       "153978   Piano Sonata: Molto moderato           0         1945       0.0866   \n",
       "\n",
       "          tempo  valence  year  \n",
       "27405   117.543   0.8610  1972  \n",
       "119814  110.689   0.3290  1956  \n",
       "155720  104.643   0.9440  1963  \n",
       "163305  101.856   0.5470  1954  \n",
       "153978  106.181   0.0448  1945  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 160k+ tracks to train our model on\n",
    "# https://www.kaggle.com/datasets/ektanegi/spotifydata-19212020?select=data.csv\n",
    "kaggle_spotify_dataset = read_csv('data.csv')\n",
    "\n",
    "kaggle_spotify_dataset.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(169909, 19)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dimensions of the dataset (rows, cols)\n",
    "kaggle_spotify_dataset.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>acousticness</th>\n",
       "      <th>danceability</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>energy</th>\n",
       "      <th>explicit</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>key</th>\n",
       "      <th>liveness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>mode</th>\n",
       "      <th>popularity</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>tempo</th>\n",
       "      <th>valence</th>\n",
       "      <th>year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>1.699090e+05</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "      <td>169909.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.493214</td>\n",
       "      <td>0.538150</td>\n",
       "      <td>2.314062e+05</td>\n",
       "      <td>0.488593</td>\n",
       "      <td>0.084863</td>\n",
       "      <td>0.161937</td>\n",
       "      <td>5.200519</td>\n",
       "      <td>0.206690</td>\n",
       "      <td>-11.370289</td>\n",
       "      <td>0.708556</td>\n",
       "      <td>31.556610</td>\n",
       "      <td>0.094058</td>\n",
       "      <td>116.948017</td>\n",
       "      <td>0.532095</td>\n",
       "      <td>1977.223231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.376627</td>\n",
       "      <td>0.175346</td>\n",
       "      <td>1.213219e+05</td>\n",
       "      <td>0.267390</td>\n",
       "      <td>0.278679</td>\n",
       "      <td>0.309329</td>\n",
       "      <td>3.515257</td>\n",
       "      <td>0.176796</td>\n",
       "      <td>5.666765</td>\n",
       "      <td>0.454429</td>\n",
       "      <td>21.582614</td>\n",
       "      <td>0.149937</td>\n",
       "      <td>30.726937</td>\n",
       "      <td>0.262408</td>\n",
       "      <td>25.593168</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.108000e+03</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-60.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1921.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.094500</td>\n",
       "      <td>0.417000</td>\n",
       "      <td>1.710400e+05</td>\n",
       "      <td>0.263000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.098400</td>\n",
       "      <td>-14.470000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>0.034900</td>\n",
       "      <td>93.516000</td>\n",
       "      <td>0.322000</td>\n",
       "      <td>1957.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.492000</td>\n",
       "      <td>0.548000</td>\n",
       "      <td>2.086000e+05</td>\n",
       "      <td>0.481000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000204</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>0.135000</td>\n",
       "      <td>-10.474000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>33.000000</td>\n",
       "      <td>0.045000</td>\n",
       "      <td>114.778000</td>\n",
       "      <td>0.544000</td>\n",
       "      <td>1978.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.888000</td>\n",
       "      <td>0.667000</td>\n",
       "      <td>2.629600e+05</td>\n",
       "      <td>0.710000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.086800</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>0.263000</td>\n",
       "      <td>-7.118000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>0.075400</td>\n",
       "      <td>135.712000</td>\n",
       "      <td>0.749000</td>\n",
       "      <td>1999.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.996000</td>\n",
       "      <td>0.988000</td>\n",
       "      <td>5.403500e+06</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.855000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>0.969000</td>\n",
       "      <td>244.091000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2020.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        acousticness   danceability   duration_ms         energy  \\\n",
       "count  169909.000000  169909.000000  1.699090e+05  169909.000000   \n",
       "mean        0.493214       0.538150  2.314062e+05       0.488593   \n",
       "std         0.376627       0.175346  1.213219e+05       0.267390   \n",
       "min         0.000000       0.000000  5.108000e+03       0.000000   \n",
       "25%         0.094500       0.417000  1.710400e+05       0.263000   \n",
       "50%         0.492000       0.548000  2.086000e+05       0.481000   \n",
       "75%         0.888000       0.667000  2.629600e+05       0.710000   \n",
       "max         0.996000       0.988000  5.403500e+06       1.000000   \n",
       "\n",
       "            explicit  instrumentalness            key       liveness  \\\n",
       "count  169909.000000     169909.000000  169909.000000  169909.000000   \n",
       "mean        0.084863          0.161937       5.200519       0.206690   \n",
       "std         0.278679          0.309329       3.515257       0.176796   \n",
       "min         0.000000          0.000000       0.000000       0.000000   \n",
       "25%         0.000000          0.000000       2.000000       0.098400   \n",
       "50%         0.000000          0.000204       5.000000       0.135000   \n",
       "75%         0.000000          0.086800       8.000000       0.263000   \n",
       "max         1.000000          1.000000      11.000000       1.000000   \n",
       "\n",
       "            loudness           mode     popularity    speechiness  \\\n",
       "count  169909.000000  169909.000000  169909.000000  169909.000000   \n",
       "mean      -11.370289       0.708556      31.556610       0.094058   \n",
       "std         5.666765       0.454429      21.582614       0.149937   \n",
       "min       -60.000000       0.000000       0.000000       0.000000   \n",
       "25%       -14.470000       0.000000      12.000000       0.034900   \n",
       "50%       -10.474000       1.000000      33.000000       0.045000   \n",
       "75%        -7.118000       1.000000      48.000000       0.075400   \n",
       "max         3.855000       1.000000     100.000000       0.969000   \n",
       "\n",
       "               tempo        valence           year  \n",
       "count  169909.000000  169909.000000  169909.000000  \n",
       "mean      116.948017       0.532095    1977.223231  \n",
       "std        30.726937       0.262408      25.593168  \n",
       "min         0.000000       0.000000    1921.000000  \n",
       "25%        93.516000       0.322000    1957.000000  \n",
       "50%       114.778000       0.544000    1978.000000  \n",
       "75%       135.712000       0.749000    1999.000000  \n",
       "max       244.091000       1.000000    2020.000000  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kaggle_spotify_dataset.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>acousticness</th>\n",
       "      <th>danceability</th>\n",
       "      <th>liveness</th>\n",
       "      <th>energy</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>speechiness</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.995</td>\n",
       "      <td>0.708</td>\n",
       "      <td>0.1510</td>\n",
       "      <td>0.1950</td>\n",
       "      <td>0.563</td>\n",
       "      <td>-12.428</td>\n",
       "      <td>0.0506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.994</td>\n",
       "      <td>0.379</td>\n",
       "      <td>0.0763</td>\n",
       "      <td>0.0135</td>\n",
       "      <td>0.901</td>\n",
       "      <td>-28.454</td>\n",
       "      <td>0.0462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.604</td>\n",
       "      <td>0.749</td>\n",
       "      <td>0.1190</td>\n",
       "      <td>0.2200</td>\n",
       "      <td>0.000</td>\n",
       "      <td>-19.924</td>\n",
       "      <td>0.9290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.995</td>\n",
       "      <td>0.781</td>\n",
       "      <td>0.1110</td>\n",
       "      <td>0.1300</td>\n",
       "      <td>0.887</td>\n",
       "      <td>-14.734</td>\n",
       "      <td>0.0926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.990</td>\n",
       "      <td>0.210</td>\n",
       "      <td>0.0980</td>\n",
       "      <td>0.2040</td>\n",
       "      <td>0.908</td>\n",
       "      <td>-16.829</td>\n",
       "      <td>0.0424</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   acousticness  danceability  liveness  energy  instrumentalness  loudness  \\\n",
       "0         0.995         0.708    0.1510  0.1950             0.563   -12.428   \n",
       "1         0.994         0.379    0.0763  0.0135             0.901   -28.454   \n",
       "2         0.604         0.749    0.1190  0.2200             0.000   -19.924   \n",
       "3         0.995         0.781    0.1110  0.1300             0.887   -14.734   \n",
       "4         0.990         0.210    0.0980  0.2040             0.908   -16.829   \n",
       "\n",
       "   speechiness  \n",
       "0       0.0506  \n",
       "1       0.0462  \n",
       "2       0.9290  \n",
       "3       0.0926  \n",
       "4       0.0424  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# X represents the features we will use to train our model\n",
    "X = kaggle_spotify_dataset[[\"acousticness\", \"danceability\", \"liveness\", \"energy\", \"instrumentalness\", \"loudness\", \"speechiness\"]]\n",
    "# All of them are in the range of [0,1] except for loudness, so we will need to rescale this feature\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = preprocessing.MinMaxScaler()\n",
    "loudness = X[[\"loudness\"]].values\n",
    "# Scale the loudness feature using Min-Max normalization\n",
    "loudness_scaled = scaler.fit_transform(loudness.reshape(-1, 1))\n",
    "X.loc[:, \"loudness\"] = loudness_scaled\n",
    "features = X.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>acousticness</th>\n",
       "      <th>danceability</th>\n",
       "      <th>liveness</th>\n",
       "      <th>energy</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>speechiness</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.995</td>\n",
       "      <td>0.708</td>\n",
       "      <td>0.1510</td>\n",
       "      <td>0.1950</td>\n",
       "      <td>0.563</td>\n",
       "      <td>0.745000</td>\n",
       "      <td>0.0506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.994</td>\n",
       "      <td>0.379</td>\n",
       "      <td>0.0763</td>\n",
       "      <td>0.0135</td>\n",
       "      <td>0.901</td>\n",
       "      <td>0.494026</td>\n",
       "      <td>0.0462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.604</td>\n",
       "      <td>0.749</td>\n",
       "      <td>0.1190</td>\n",
       "      <td>0.2200</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.627609</td>\n",
       "      <td>0.9290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.995</td>\n",
       "      <td>0.781</td>\n",
       "      <td>0.1110</td>\n",
       "      <td>0.1300</td>\n",
       "      <td>0.887</td>\n",
       "      <td>0.708887</td>\n",
       "      <td>0.0926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.990</td>\n",
       "      <td>0.210</td>\n",
       "      <td>0.0980</td>\n",
       "      <td>0.2040</td>\n",
       "      <td>0.908</td>\n",
       "      <td>0.676079</td>\n",
       "      <td>0.0424</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   acousticness  danceability  liveness  energy  instrumentalness  loudness  \\\n",
       "0         0.995         0.708    0.1510  0.1950             0.563  0.745000   \n",
       "1         0.994         0.379    0.0763  0.0135             0.901  0.494026   \n",
       "2         0.604         0.749    0.1190  0.2200             0.000  0.627609   \n",
       "3         0.995         0.781    0.1110  0.1300             0.887  0.708887   \n",
       "4         0.990         0.210    0.0980  0.2040             0.908  0.676079   \n",
       "\n",
       "   speechiness  \n",
       "0       0.0506  \n",
       "1       0.0462  \n",
       "2       0.9290  \n",
       "3       0.0926  \n",
       "4       0.0424  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Choosing a value for K (# clusters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model with 1 clusters trained\n",
      "Model with 2 clusters trained\n",
      "Model with 3 clusters trained\n",
      "Model with 4 clusters trained\n",
      "Model with 5 clusters trained\n",
      "Model with 6 clusters trained\n",
      "Model with 7 clusters trained\n",
      "Model with 8 clusters trained\n",
      "Model with 9 clusters trained\n",
      "Model with 10 clusters trained\n"
     ]
    }
   ],
   "source": [
    "# Train a KMeans model for k=1 to k=10 clusters and keep track of the sum of squared distances for each k\n",
    "sum_of_squared_distances = []\n",
    "for k in range(1, 11):\n",
    "    model = KMeans(n_clusters=k)\n",
    "    model = model.fit(features)\n",
    "    sum_of_squared_distances.append(model.inertia_)\n",
    "    print(f\"Model with {k} clusters trained\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 5))\n",
    "plt.plot(range(1, 11), sum_of_squared_distances, 'bx-')\n",
    "plt.xlabel('k')\n",
    "plt.ylabel('Sum of squared distances')\n",
    "plt.title('Elbow Method For Optimal k')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train the KMeans model with k=5 clusters\n",
    "model = KMeans(n_clusters=5)\n",
    "model = model.fit(features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 2, 1, ..., 1, 3, 3])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Predict the cluster for each song\n",
    "predictions = model.predict(features)\n",
    "predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>acousticness</th>\n",
       "      <th>artists</th>\n",
       "      <th>danceability</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>energy</th>\n",
       "      <th>explicit</th>\n",
       "      <th>id</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>key</th>\n",
       "      <th>liveness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>mode</th>\n",
       "      <th>name</th>\n",
       "      <th>popularity</th>\n",
       "      <th>release_date</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>tempo</th>\n",
       "      <th>valence</th>\n",
       "      <th>year</th>\n",
       "      <th>cluster</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>51809</th>\n",
       "      <td>0.16000</td>\n",
       "      <td>['Nazia Hassan']</td>\n",
       "      <td>0.672</td>\n",
       "      <td>321080</td>\n",
       "      <td>0.5560</td>\n",
       "      <td>0</td>\n",
       "      <td>6dxbXni1MNwzNSJ3LNYh7h</td>\n",
       "      <td>0.000226</td>\n",
       "      <td>4</td>\n",
       "      <td>0.274</td>\n",
       "      <td>-13.436</td>\n",
       "      <td>0</td>\n",
       "      <td>Boom Boom</td>\n",
       "      <td>31</td>\n",
       "      <td>1982-09-01</td>\n",
       "      <td>0.0281</td>\n",
       "      <td>129.689</td>\n",
       "      <td>0.904</td>\n",
       "      <td>1982</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60758</th>\n",
       "      <td>0.00672</td>\n",
       "      <td>['Rancid']</td>\n",
       "      <td>0.452</td>\n",
       "      <td>148333</td>\n",
       "      <td>0.9880</td>\n",
       "      <td>0</td>\n",
       "      <td>18gBnrdlaLh0uUohNfWqlk</td>\n",
       "      <td>0.000136</td>\n",
       "      <td>0</td>\n",
       "      <td>0.404</td>\n",
       "      <td>-5.634</td>\n",
       "      <td>1</td>\n",
       "      <td>You Don't Care Nothin'</td>\n",
       "      <td>37</td>\n",
       "      <td>1995-08-15</td>\n",
       "      <td>0.0722</td>\n",
       "      <td>167.980</td>\n",
       "      <td>0.673</td>\n",
       "      <td>1995</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101072</th>\n",
       "      <td>0.72000</td>\n",
       "      <td>['Francisco \"Charro\" Avitia']</td>\n",
       "      <td>0.374</td>\n",
       "      <td>213893</td>\n",
       "      <td>0.3610</td>\n",
       "      <td>0</td>\n",
       "      <td>6SAtOJEknkFihBbPuf5k3H</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>9</td>\n",
       "      <td>0.203</td>\n",
       "      <td>-5.690</td>\n",
       "      <td>1</td>\n",
       "      <td>Corrido de Santa Amalia</td>\n",
       "      <td>32</td>\n",
       "      <td>1958-01-12</td>\n",
       "      <td>0.0361</td>\n",
       "      <td>183.286</td>\n",
       "      <td>0.611</td>\n",
       "      <td>1958</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27522</th>\n",
       "      <td>0.64700</td>\n",
       "      <td>['Patty Pravo']</td>\n",
       "      <td>0.356</td>\n",
       "      <td>288093</td>\n",
       "      <td>0.5450</td>\n",
       "      <td>0</td>\n",
       "      <td>7n9YiIfoZwYOzvUc2jrrpp</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4</td>\n",
       "      <td>0.485</td>\n",
       "      <td>-6.334</td>\n",
       "      <td>1</td>\n",
       "      <td>Pazza idea</td>\n",
       "      <td>48</td>\n",
       "      <td>1973-01-14</td>\n",
       "      <td>0.0484</td>\n",
       "      <td>153.496</td>\n",
       "      <td>0.550</td>\n",
       "      <td>1973</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91338</th>\n",
       "      <td>0.95300</td>\n",
       "      <td>['Vince Guaraldi Trio']</td>\n",
       "      <td>0.452</td>\n",
       "      <td>295093</td>\n",
       "      <td>0.0556</td>\n",
       "      <td>0</td>\n",
       "      <td>2eZCAJoigZpdSJY9LK3InG</td>\n",
       "      <td>0.867000</td>\n",
       "      <td>5</td>\n",
       "      <td>0.102</td>\n",
       "      <td>-24.216</td>\n",
       "      <td>0</td>\n",
       "      <td>Django</td>\n",
       "      <td>32</td>\n",
       "      <td>1956</td>\n",
       "      <td>0.0488</td>\n",
       "      <td>90.668</td>\n",
       "      <td>0.196</td>\n",
       "      <td>1956</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        acousticness                        artists  danceability  \\\n",
       "51809        0.16000               ['Nazia Hassan']         0.672   \n",
       "60758        0.00672                     ['Rancid']         0.452   \n",
       "101072       0.72000  ['Francisco \"Charro\" Avitia']         0.374   \n",
       "27522        0.64700                ['Patty Pravo']         0.356   \n",
       "91338        0.95300        ['Vince Guaraldi Trio']         0.452   \n",
       "\n",
       "        duration_ms  energy  explicit                      id  \\\n",
       "51809        321080  0.5560         0  6dxbXni1MNwzNSJ3LNYh7h   \n",
       "60758        148333  0.9880         0  18gBnrdlaLh0uUohNfWqlk   \n",
       "101072       213893  0.3610         0  6SAtOJEknkFihBbPuf5k3H   \n",
       "27522        288093  0.5450         0  7n9YiIfoZwYOzvUc2jrrpp   \n",
       "91338        295093  0.0556         0  2eZCAJoigZpdSJY9LK3InG   \n",
       "\n",
       "        instrumentalness  key  liveness  loudness  mode  \\\n",
       "51809           0.000226    4     0.274   -13.436     0   \n",
       "60758           0.000136    0     0.404    -5.634     1   \n",
       "101072          0.000005    9     0.203    -5.690     1   \n",
       "27522           0.000000    4     0.485    -6.334     1   \n",
       "91338           0.867000    5     0.102   -24.216     0   \n",
       "\n",
       "                           name  popularity release_date  speechiness  \\\n",
       "51809                 Boom Boom          31   1982-09-01       0.0281   \n",
       "60758    You Don't Care Nothin'          37   1995-08-15       0.0722   \n",
       "101072  Corrido de Santa Amalia          32   1958-01-12       0.0361   \n",
       "27522                Pazza idea          48   1973-01-14       0.0484   \n",
       "91338                    Django          32         1956       0.0488   \n",
       "\n",
       "          tempo  valence  year  cluster  \n",
       "51809   129.689    0.904  1982        3  \n",
       "60758   167.980    0.673  1995        3  \n",
       "101072  183.286    0.611  1958        0  \n",
       "27522   153.496    0.550  1973        1  \n",
       "91338    90.668    0.196  1956        2  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kaggle_spotify_dataset['cluster'] = predictions\n",
    "kaggle_spotify_dataset.sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>cluster</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.658976</td>\n",
       "      <td>0.114121</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.954404</td>\n",
       "      <td>0.423143</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.128846</td>\n",
       "      <td>-0.345963</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.807799</td>\n",
       "      <td>0.378912</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.858238</td>\n",
       "      <td>0.496279</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0         1  cluster\n",
       "0  0.658976  0.114121        2\n",
       "1  0.954404  0.423143        2\n",
       "2  0.128846 -0.345963        1\n",
       "3  0.807799  0.378912        2\n",
       "4  0.858238  0.496279        2"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca = decomposition.PCA(n_components=2)\n",
    "pca_result = pca.fit_transform(features)\n",
    "df = pd.DataFrame(pca_result)\n",
    "df['cluster'] = predictions\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(10, 5))\n",
    "ax = fig.add_subplot(111)\n",
    "ax.set_xlabel('x')\n",
    "ax.set_ylabel('y')\n",
    "scatter = ax.scatter(df[0], df[1], c=df['cluster'].values, s=50)\n",
    "plt.colorbar(scatter)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training different classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_classify = kaggle_spotify_dataset[[\"acousticness\", \"danceability\", \"liveness\",\"energy\", \"instrumentalness\", \"loudness\", \"speechiness\", \"cluster\"]]\n",
    "X = df_classify.iloc[:, :-1].values\n",
    "Y = df_classify.iloc[:, -1].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split the dataset into training and testing sets (80% training, 20% testing)\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scale the 5th feature (loudness) using Min-Max normalization\n",
    "X_train[:, 5] = scaler.fit_transform(X_train[:, 5].reshape(-1, 1)).reshape(-1)\n",
    "X_test[:, 5] = scaler.transform(X_test[:, 5].reshape(-1, 1)).reshape(-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2:0.9743393561297158\n",
      "3:0.980666235065623\n",
      "4:0.9798422694367607\n",
      "5:0.9827850038255547\n",
      "6:0.9820493202283562\n",
      "7:0.9838149608616327\n",
      "8:0.9839326702371843\n",
      "9:0.9844917897710552\n"
     ]
    }
   ],
   "source": [
    "# Train a KNN classifier for k=2 to k=9 neighbors and keep track of the accuracy for each k\n",
    "acc=[]\n",
    "for i in range(2,10):\n",
    "  classifier = KNeighborsClassifier(n_neighbors = i, metric = 'minkowski', p = 2)\n",
    "  classifier.fit(X_train, Y_train)\n",
    "  Y_pred = classifier.predict(X_test)\n",
    "  acc.append(accuracy_score(Y_test, Y_pred))\n",
    "  print(f'{i}:{acc[-1]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5:0.9860808663410041\n",
      "6:0.9858454475899006\n",
      "7:0.9871402507209699\n",
      "8:0.987081396033194\n",
      "9:0.988081925725384\n",
      "10:0.9877876522865047\n",
      "11:0.9884644811959272\n",
      "12:0.9887293272909187\n",
      "13:0.9892295921370137\n"
     ]
    }
   ],
   "source": [
    "# Train a Random Forest classifier for n=5 to n=13 trees and keep track of the accuracy for each n\n",
    "acc=[]\n",
    "for i in range(5,14):\n",
    "  classifier = RandomForestClassifier(n_estimators = i, criterion = 'entropy', random_state = 0)\n",
    "  classifier.fit(X_train, Y_train)\n",
    "  Y_pred = classifier.predict(X_test)\n",
    "  acc.append(accuracy_score(Y_test, Y_pred))\n",
    "  print(f'{i}:{acc[-1]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13:0.9333470660938143\n"
     ]
    }
   ],
   "source": [
    "# Train a Naive Bayes classifier\n",
    "acc=[]\n",
    "\n",
    "classifier = GaussianNB()\n",
    "classifier.fit(X_train, Y_train)\n",
    "Y_pred = classifier.predict(X_test)\n",
    "acc.append(accuracy_score(Y_test, Y_pred))\n",
    "print(f'{i}:{acc[-1]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "linear:0.9975869578011889\n",
      "poly:0.9974692484256371\n",
      "rbf:0.9971749749867577\n",
      "sigmoid:0.6243599552704373\n"
     ]
    }
   ],
   "source": [
    "# Train a SVM classifier for each kernel and keep track of the accuracy for each kernel\n",
    "acc=[]\n",
    "for i in ['linear', 'poly', 'rbf', 'sigmoid']:\n",
    "  classifier = SVC(kernel = i, random_state = 0)\n",
    "  classifier.fit(X_train, Y_train)\n",
    "  Y_pred = classifier.predict(X_test)\n",
    "  acc.append(accuracy_score(Y_test, Y_pred))\n",
    "  print(f'{i}:{acc[-1]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier = SVC(kernel ='poly', random_state = 0)\n",
    "classifier.fit(X_train, Y_train)\n",
    "Y_pred = classifier.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 8775    10     5     0     0]\n",
      " [   10  7435     1    22     2]\n",
      " [    2     0  4518     0     5]\n",
      " [    0    16     0 11681     5]\n",
      " [    0     3     2     3  1487]]\n"
     ]
    }
   ],
   "source": [
    "cm = confusion_matrix(Y_test, Y_pred)\n",
    "print(cm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x700 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_cm = pd.DataFrame(cm, range(5), range(5))\n",
    "plt.figure(figsize=(10,7))\n",
    "sns.set_theme(font_scale=1.4)\n",
    "sns.heatmap(df_cm, annot=True, annot_kws={\"size\": 16})\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cluster description"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[acousticness         0.878883\n",
       " danceability         0.489093\n",
       " liveness             0.208687\n",
       " energy               0.259463\n",
       " instrumentalness     0.037162\n",
       " loudness           -14.020297\n",
       " speechiness          0.086344\n",
       " cluster              0.000000\n",
       " Name: mean, dtype: float64,\n",
       " acousticness         0.473721\n",
       " danceability         0.602656\n",
       " liveness             0.208551\n",
       " energy               0.493430\n",
       " instrumentalness     0.015160\n",
       " loudness           -10.754410\n",
       " speechiness          0.140079\n",
       " cluster              1.000000\n",
       " Name: mean, dtype: float64,\n",
       " acousticness         0.929106\n",
       " danceability         0.418680\n",
       " liveness             0.178713\n",
       " energy               0.225010\n",
       " instrumentalness     0.819364\n",
       " loudness           -17.896771\n",
       " speechiness          0.060241\n",
       " cluster              2.000000\n",
       " Name: mean, dtype: float64,\n",
       " acousticness        0.086389\n",
       " danceability        0.586584\n",
       " liveness            0.212100\n",
       " energy              0.737803\n",
       " instrumentalness    0.017393\n",
       " loudness           -7.323537\n",
       " speechiness         0.088126\n",
       " cluster             3.000000\n",
       " Name: mean, dtype: float64,\n",
       " acousticness         0.150658\n",
       " danceability         0.497154\n",
       " liveness             0.229061\n",
       " energy               0.681833\n",
       " instrumentalness     0.716033\n",
       " loudness           -10.255339\n",
       " speechiness          0.060856\n",
       " cluster              4.000000\n",
       " Name: mean, dtype: float64]"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = kaggle_spotify_dataset[[\"id\",\"acousticness\", \"danceability\", \"liveness\",\"energy\", \"instrumentalness\", \"loudness\", \"speechiness\", \"cluster\"]]\n",
    "\n",
    "cluster_info = []\n",
    "for i in range(5):\n",
    "  cluster_info.append(df[df['cluster'] == i].describe().iloc[1,:])\n",
    "\n",
    "cluster_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classifying my liked songs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('config.json') as f:\n",
    "    config = json.load(f)\n",
    "\n",
    "cid = config['cid']\n",
    "secret = config['secret']\n",
    "username = config['username']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Avaliable scopes: https://developer.spotify.com/web-api/using-scopes/\n",
    "scope = 'user-library-read playlist-modify-public playlist-read-private user-top-read playlist-modify-private user-read-currently-playing'\n",
    "redirect_uri='https://example.com/callback'\n",
    "client_credentials_manager = spotipy.oauth2.SpotifyOAuth(scope=scope, username=username,client_id=cid, client_secret=secret, redirect_uri=redirect_uri, cache_path='./tokens.txt') \n",
    "sp = spotipy.Spotify(client_credentials_manager=client_credentials_manager)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "offset = 0\n",
    "songs = []\n",
    "names = []\n",
    "ids = []\n",
    "\n",
    "# Retrieve my liked songs in batches of 50 until all songs are fetched\n",
    "while True:\n",
    "    # Fetch tracks from Spotify API\n",
    "    content = sp.current_user_saved_tracks(limit=50, offset=offset)\n",
    "    songs += content['items']\n",
    "\n",
    "    # If there are more tracks to fetch, increase the offset\n",
    "    if content['next'] is not None:\n",
    "        offset += 50\n",
    "    else:\n",
    "        break\n",
    "\n",
    "# Extract the name and id of each song\n",
    "for i in songs:\n",
    "    names.append(i['track']['name'])\n",
    "    ids.append(i['track']['id'])\n",
    "\n",
    "index=0\n",
    "audio_features=[]\n",
    "\n",
    "# Retrieve the audio features of my liked songs in batches of 50\n",
    "while index < len(ids):\n",
    "    # Fetch audio features from Spotify API\n",
    "    audio_features += sp.audio_features(ids[index:index + 50])\n",
    "    index += 50\n",
    "\n",
    "# Create a list of lists with the audio features of each song\n",
    "features_list = []\n",
    "for features in audio_features:\n",
    "    features_list.append([features['acousticness'], features['danceability'], features['liveness'], features['energy'], features['instrumentalness'], features['loudness'], features['speechiness']])\n",
    "\n",
    "# Create a dataframe from the list of audio features with track IDs as index\n",
    "mydf = pd.DataFrame(features_list, columns=[\"acousticness\", \"danceability\", \"liveness\",\"energy\", \"instrumentalness\", \"loudness\", \"speechiness\"],index=ids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2182"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Sanity check\n",
    "len(songs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>acousticness</th>\n",
       "      <th>danceability</th>\n",
       "      <th>liveness</th>\n",
       "      <th>energy</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>speechiness</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0JiY190vktuhSGN6aqJdrt</th>\n",
       "      <td>0.000334</td>\n",
       "      <td>0.534</td>\n",
       "      <td>0.241</td>\n",
       "      <td>0.870</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-3.078</td>\n",
       "      <td>0.0425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1QV6tiMFM6fSOKOGLMHYYg</th>\n",
       "      <td>0.118000</td>\n",
       "      <td>0.851</td>\n",
       "      <td>0.121</td>\n",
       "      <td>0.806</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>-4.620</td>\n",
       "      <td>0.0787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2ygMBIctKIAfbEBcT9065L</th>\n",
       "      <td>0.009370</td>\n",
       "      <td>0.648</td>\n",
       "      <td>0.752</td>\n",
       "      <td>0.931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-3.150</td>\n",
       "      <td>0.1810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4wH4dJgrsxONID6KS2tDQM</th>\n",
       "      <td>0.026100</td>\n",
       "      <td>0.796</td>\n",
       "      <td>0.121</td>\n",
       "      <td>0.777</td>\n",
       "      <td>0.000358</td>\n",
       "      <td>-4.810</td>\n",
       "      <td>0.0397</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6I9VzXrHxO9rA9A5euc8Ak</th>\n",
       "      <td>0.024900</td>\n",
       "      <td>0.774</td>\n",
       "      <td>0.242</td>\n",
       "      <td>0.838</td>\n",
       "      <td>0.025000</td>\n",
       "      <td>-3.914</td>\n",
       "      <td>0.1140</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        acousticness  danceability  liveness  energy  \\\n",
       "0JiY190vktuhSGN6aqJdrt      0.000334         0.534     0.241   0.870   \n",
       "1QV6tiMFM6fSOKOGLMHYYg      0.118000         0.851     0.121   0.806   \n",
       "2ygMBIctKIAfbEBcT9065L      0.009370         0.648     0.752   0.931   \n",
       "4wH4dJgrsxONID6KS2tDQM      0.026100         0.796     0.121   0.777   \n",
       "6I9VzXrHxO9rA9A5euc8Ak      0.024900         0.774     0.242   0.838   \n",
       "\n",
       "                        instrumentalness  loudness  speechiness  \n",
       "0JiY190vktuhSGN6aqJdrt          0.000000    -3.078       0.0425  \n",
       "1QV6tiMFM6fSOKOGLMHYYg          0.000002    -4.620       0.0787  \n",
       "2ygMBIctKIAfbEBcT9065L          0.000000    -3.150       0.1810  \n",
       "4wH4dJgrsxONID6KS2tDQM          0.000358    -4.810       0.0397  \n",
       "6I9VzXrHxO9rA9A5euc8Ak          0.025000    -3.914       0.1140  "
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mydf.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>acousticness</th>\n",
       "      <th>danceability</th>\n",
       "      <th>liveness</th>\n",
       "      <th>energy</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>cluster</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0JiY190vktuhSGN6aqJdrt</th>\n",
       "      <td>0.000334</td>\n",
       "      <td>0.534</td>\n",
       "      <td>0.241</td>\n",
       "      <td>0.870</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.928741</td>\n",
       "      <td>0.0425</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1QV6tiMFM6fSOKOGLMHYYg</th>\n",
       "      <td>0.118000</td>\n",
       "      <td>0.851</td>\n",
       "      <td>0.121</td>\n",
       "      <td>0.806</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.886053</td>\n",
       "      <td>0.0787</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2ygMBIctKIAfbEBcT9065L</th>\n",
       "      <td>0.009370</td>\n",
       "      <td>0.648</td>\n",
       "      <td>0.752</td>\n",
       "      <td>0.931</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.926748</td>\n",
       "      <td>0.1810</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4wH4dJgrsxONID6KS2tDQM</th>\n",
       "      <td>0.026100</td>\n",
       "      <td>0.796</td>\n",
       "      <td>0.121</td>\n",
       "      <td>0.777</td>\n",
       "      <td>0.000358</td>\n",
       "      <td>0.880793</td>\n",
       "      <td>0.0397</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6I9VzXrHxO9rA9A5euc8Ak</th>\n",
       "      <td>0.024900</td>\n",
       "      <td>0.774</td>\n",
       "      <td>0.242</td>\n",
       "      <td>0.838</td>\n",
       "      <td>0.025000</td>\n",
       "      <td>0.905598</td>\n",
       "      <td>0.1140</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        acousticness  danceability  liveness  energy  \\\n",
       "0JiY190vktuhSGN6aqJdrt      0.000334         0.534     0.241   0.870   \n",
       "1QV6tiMFM6fSOKOGLMHYYg      0.118000         0.851     0.121   0.806   \n",
       "2ygMBIctKIAfbEBcT9065L      0.009370         0.648     0.752   0.931   \n",
       "4wH4dJgrsxONID6KS2tDQM      0.026100         0.796     0.121   0.777   \n",
       "6I9VzXrHxO9rA9A5euc8Ak      0.024900         0.774     0.242   0.838   \n",
       "\n",
       "                        instrumentalness  loudness  speechiness  cluster  \n",
       "0JiY190vktuhSGN6aqJdrt          0.000000  0.928741       0.0425        3  \n",
       "1QV6tiMFM6fSOKOGLMHYYg          0.000002  0.886053       0.0787        3  \n",
       "2ygMBIctKIAfbEBcT9065L          0.000000  0.926748       0.1810        3  \n",
       "4wH4dJgrsxONID6KS2tDQM          0.000358  0.880793       0.0397        3  \n",
       "6I9VzXrHxO9rA9A5euc8Ak          0.025000  0.905598       0.1140        3  "
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loudness=mydf['loudness'].values\n",
    "loudness_scaled=scaler.fit_transform(loudness.reshape(-1,1))\n",
    "mydf['loudness'] = loudness_scaled\n",
    "\n",
    "features = mydf.values\n",
    "predictions = classifier.predict(features)\n",
    "mydf['cluster'] = predictions\n",
    "mydf.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2182, 8)"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mydf.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Combine the index (track IDs) and the cluster predictions into a list of tuples\n",
    "clustered_songs = list(zip(mydf.index,mydf.iloc[:,-1]))\n",
    "\n",
    "# Sort the songs into lists based on their cluster\n",
    "sorted_songs=[[],[],[],[],[]]\n",
    "for i in range(len(clustered_songs)):\n",
    "    sorted_songs[clustered_songs[i][1]].append(clustered_songs[i][0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "204\n",
      "502\n",
      "21\n",
      "1372\n",
      "83\n"
     ]
    }
   ],
   "source": [
    "# Sanity check\n",
    "print(len(sorted_songs[0]))\n",
    "print(len(sorted_songs[1]))\n",
    "print(len(sorted_songs[2]))\n",
    "print(len(sorted_songs[3]))\n",
    "print(len(sorted_songs[4]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create 5 playlists with the sorted songs\n",
    "playlist_names = ['🐸','🪐','🦊','🐳','🦜']\n",
    "playlists = []\n",
    "for i in range(5):\n",
    "    # Create a playlist for each cluster\n",
    "    playlist_created = sp.user_playlist_create(username, name=playlist_names[i], public=False,description='')\n",
    "    for j in range(0, len(sorted_songs[i]), 100):\n",
    "        # Add songs to the playlist in batches of 100\n",
    "        sp.user_playlist_add_tracks(username,playlist_created['id'],sorted_songs[i][j:j + 100])\n",
    "        playlists.append(playlist_created)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "my saved songs:\n",
      "0 0.09349220898258478\n",
      "1 0.23006416131989002\n",
      "2 0.009624197983501375\n",
      "3 0.6287809349220899\n",
      "4 0.038038496791934\n"
     ]
    }
   ],
   "source": [
    "# Count the number of songs in each cluster\n",
    "len_clusters = []\n",
    "for i in range(5):\n",
    "    len_clusters.append(len(mydf[mydf['cluster']==i].values))\n",
    "\n",
    "print('my saved songs:')\n",
    "\n",
    "# Calculate the percentage of songs in each cluster\n",
    "pers = []\n",
    "for i in range(5):\n",
    "    pers.append(float(len_clusters[i]/sum(len_clusters)))\n",
    "    print(i,float(len_clusters[i]/sum(len_clusters)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sources"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://github.com/chingisooinar/Clustering-my-songs-on-Spotify/blob/main/Clustering_my_songs_on_Spotify.ipynb\n",
    "\n",
    "https://towardsdatascience.com/cluster-your-liked-songs-on-spotify-into-playlists-of-similar-songs-66a244ba297e"
   ]
  }
 ],
 "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": "3.11.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}