Make_Your_Own_Playlist_of_Recs-with_feature_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creates a playlist of best recommended tracks per track in a playlist\n",
    "\n",
    "# install Spotipy first\n",
    "#!pip install Spotipy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import spotipy\n",
    "from spotipy.oauth2 import SpotifyClientCredentials\n",
    "import spotipy.util as util\n",
    "\n",
    "\n",
    "cid =' ' # Client ID; copy this from your app created on beta.developer.spotify.com\n",
    "secret = ' ' # Client Secret; copy this from your app\n",
    "username = ' ' # Your Spotify username\n",
    "\n",
    "#for avaliable scopes see https://developer.spotify.com/web-api/using-scopes/\n",
    "scope = 'user-library-read playlist-modify-public playlist-read-private'\n",
    "\n",
    "redirect_uri='https://beta.developer.spotify.com/dashboard/applications/ef60e9f2d37b4913963c7e7e9c572c96'\n",
    "\n",
    "client_credentials_manager = SpotifyClientCredentials(client_id=cid, client_secret=secret) \n",
    "\n",
    "sp = spotipy.Spotify(client_credentials_manager=client_credentials_manager)\n",
    "\n",
    "token = util.prompt_for_user_token(username, scope, cid, secret, redirect_uri)\n",
    "\n",
    "if token:\n",
    "    sp = spotipy.Spotify(auth=token)\n",
    "else:\n",
    "    print(\"Can't get token for\", username)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Create a dataframe of your playlist including tracks' names and audio features\n",
    "import pandas as pd\n",
    "\n",
    "sourcePlaylistID = '1fCOovfyVaAbUiPlqtRF09'\n",
    "sourcePlaylist = sp.user_playlist(username, sourcePlaylistID);\n",
    "tracks = sourcePlaylist[\"tracks\"];\n",
    "songs = tracks[\"items\"];\n",
    "\n",
    "track_ids = []\n",
    "track_names = []\n",
    "\n",
    "for i in range(0, len(songs)):\n",
    "    if songs[i]['track']['id'] != None: # Removes the local tracks in your playlist if there is any\n",
    "        track_ids.append(songs[i]['track']['id'])\n",
    "        track_names.append(songs[i]['track']['name'])\n",
    "\n",
    "features = []\n",
    "for i in range(0,len(track_ids)):\n",
    "    audio_features = sp.audio_features(track_ids[i])\n",
    "    for track in audio_features:\n",
    "        features.append(track)\n",
    "        \n",
    "playlist_df = pd.DataFrame(features, index = track_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "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>analysis_url</th>\n",
       "      <th>danceability</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>energy</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>speechiness</th>\n",
       "      <th>tempo</th>\n",
       "      <th>time_signature</th>\n",
       "      <th>track_href</th>\n",
       "      <th>type</th>\n",
       "      <th>uri</th>\n",
       "      <th>valence</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>The Bluest Blues (feat. George Harrison)</th>\n",
       "      <td>0.5180</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/3d5e...</td>\n",
       "      <td>0.404</td>\n",
       "      <td>447240</td>\n",
       "      <td>0.331</td>\n",
       "      <td>3d5e9kJDkgc7Jum8d8xtrl</td>\n",
       "      <td>0.353000</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0786</td>\n",
       "      <td>-11.487</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0245</td>\n",
       "      <td>73.904</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/3d5e9kJDkgc7...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:3d5e9kJDkgc7Jum8d8xtrl</td>\n",
       "      <td>0.2470</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bird of Paradise</th>\n",
       "      <td>0.5960</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/0X7R...</td>\n",
       "      <td>0.471</td>\n",
       "      <td>281467</td>\n",
       "      <td>0.437</td>\n",
       "      <td>0X7RPlCMbifRgyg1RpsR0K</td>\n",
       "      <td>0.641000</td>\n",
       "      <td>1</td>\n",
       "      <td>0.2780</td>\n",
       "      <td>-11.396</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0370</td>\n",
       "      <td>144.333</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/0X7RPlCMbifR...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:0X7RPlCMbifRgyg1RpsR0K</td>\n",
       "      <td>0.3370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The Thrill Is Gone</th>\n",
       "      <td>0.0934</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/1xW9...</td>\n",
       "      <td>0.692</td>\n",
       "      <td>301133</td>\n",
       "      <td>0.493</td>\n",
       "      <td>1xW9HHRQe8MMGF8SBNgkBY</td>\n",
       "      <td>0.001660</td>\n",
       "      <td>9</td>\n",
       "      <td>0.1280</td>\n",
       "      <td>-9.662</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0325</td>\n",
       "      <td>86.644</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/1xW9HHRQe8MM...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:1xW9HHRQe8MMGF8SBNgkBY</td>\n",
       "      <td>0.2800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Times Makes Two</th>\n",
       "      <td>0.6430</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/5LJF...</td>\n",
       "      <td>0.641</td>\n",
       "      <td>325987</td>\n",
       "      <td>0.537</td>\n",
       "      <td>5LJFs3O1BFipvKh2wZHNMs</td>\n",
       "      <td>0.000003</td>\n",
       "      <td>10</td>\n",
       "      <td>0.0528</td>\n",
       "      <td>-6.921</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0342</td>\n",
       "      <td>112.085</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/5LJFs3O1BFip...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:5LJFs3O1BFipvKh2wZHNMs</td>\n",
       "      <td>0.3260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Riding the Blues</th>\n",
       "      <td>0.5050</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/6r8w...</td>\n",
       "      <td>0.438</td>\n",
       "      <td>226987</td>\n",
       "      <td>0.432</td>\n",
       "      <td>6r8w9NhDqcMY3l8SicsaDt</td>\n",
       "      <td>0.005210</td>\n",
       "      <td>4</td>\n",
       "      <td>0.1440</td>\n",
       "      <td>-7.931</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0318</td>\n",
       "      <td>127.104</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/6r8w9NhDqcMY...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:6r8w9NhDqcMY3l8SicsaDt</td>\n",
       "      <td>0.0904</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          acousticness  \\\n",
       "The Bluest Blues (feat. George Harrison)        0.5180   \n",
       "Bird of Paradise                                0.5960   \n",
       "The Thrill Is Gone                              0.0934   \n",
       "Times Makes Two                                 0.6430   \n",
       "Riding the Blues                                0.5050   \n",
       "\n",
       "                                                                               analysis_url  \\\n",
       "The Bluest Blues (feat. George Harrison)  https://api.spotify.com/v1/audio-analysis/3d5e...   \n",
       "Bird of Paradise                          https://api.spotify.com/v1/audio-analysis/0X7R...   \n",
       "The Thrill Is Gone                        https://api.spotify.com/v1/audio-analysis/1xW9...   \n",
       "Times Makes Two                           https://api.spotify.com/v1/audio-analysis/5LJF...   \n",
       "Riding the Blues                          https://api.spotify.com/v1/audio-analysis/6r8w...   \n",
       "\n",
       "                                          danceability  duration_ms  energy  \\\n",
       "The Bluest Blues (feat. George Harrison)         0.404       447240   0.331   \n",
       "Bird of Paradise                                 0.471       281467   0.437   \n",
       "The Thrill Is Gone                               0.692       301133   0.493   \n",
       "Times Makes Two                                  0.641       325987   0.537   \n",
       "Riding the Blues                                 0.438       226987   0.432   \n",
       "\n",
       "                                                              id  \\\n",
       "The Bluest Blues (feat. George Harrison)  3d5e9kJDkgc7Jum8d8xtrl   \n",
       "Bird of Paradise                          0X7RPlCMbifRgyg1RpsR0K   \n",
       "The Thrill Is Gone                        1xW9HHRQe8MMGF8SBNgkBY   \n",
       "Times Makes Two                           5LJFs3O1BFipvKh2wZHNMs   \n",
       "Riding the Blues                          6r8w9NhDqcMY3l8SicsaDt   \n",
       "\n",
       "                                          instrumentalness  key  liveness  \\\n",
       "The Bluest Blues (feat. George Harrison)          0.353000    9    0.0786   \n",
       "Bird of Paradise                                  0.641000    1    0.2780   \n",
       "The Thrill Is Gone                                0.001660    9    0.1280   \n",
       "Times Makes Two                                   0.000003   10    0.0528   \n",
       "Riding the Blues                                  0.005210    4    0.1440   \n",
       "\n",
       "                                          loudness  mode  speechiness  \\\n",
       "The Bluest Blues (feat. George Harrison)   -11.487     0       0.0245   \n",
       "Bird of Paradise                           -11.396     1       0.0370   \n",
       "The Thrill Is Gone                          -9.662     0       0.0325   \n",
       "Times Makes Two                             -6.921     0       0.0342   \n",
       "Riding the Blues                            -7.931     0       0.0318   \n",
       "\n",
       "                                            tempo  time_signature  \\\n",
       "The Bluest Blues (feat. George Harrison)   73.904               4   \n",
       "Bird of Paradise                          144.333               4   \n",
       "The Thrill Is Gone                         86.644               4   \n",
       "Times Makes Two                           112.085               4   \n",
       "Riding the Blues                          127.104               4   \n",
       "\n",
       "                                                                                 track_href  \\\n",
       "The Bluest Blues (feat. George Harrison)  https://api.spotify.com/v1/tracks/3d5e9kJDkgc7...   \n",
       "Bird of Paradise                          https://api.spotify.com/v1/tracks/0X7RPlCMbifR...   \n",
       "The Thrill Is Gone                        https://api.spotify.com/v1/tracks/1xW9HHRQe8MM...   \n",
       "Times Makes Two                           https://api.spotify.com/v1/tracks/5LJFs3O1BFip...   \n",
       "Riding the Blues                          https://api.spotify.com/v1/tracks/6r8w9NhDqcMY...   \n",
       "\n",
       "                                                    type  \\\n",
       "The Bluest Blues (feat. George Harrison)  audio_features   \n",
       "Bird of Paradise                          audio_features   \n",
       "The Thrill Is Gone                        audio_features   \n",
       "Times Makes Two                           audio_features   \n",
       "Riding the Blues                          audio_features   \n",
       "\n",
       "                                                                           uri  \\\n",
       "The Bluest Blues (feat. George Harrison)  spotify:track:3d5e9kJDkgc7Jum8d8xtrl   \n",
       "Bird of Paradise                          spotify:track:0X7RPlCMbifRgyg1RpsR0K   \n",
       "The Thrill Is Gone                        spotify:track:1xW9HHRQe8MMGF8SBNgkBY   \n",
       "Times Makes Two                           spotify:track:5LJFs3O1BFipvKh2wZHNMs   \n",
       "Riding the Blues                          spotify:track:6r8w9NhDqcMY3l8SicsaDt   \n",
       "\n",
       "                                          valence  \n",
       "The Bluest Blues (feat. George Harrison)   0.2470  \n",
       "Bird of Paradise                           0.3370  \n",
       "The Thrill Is Gone                         0.2800  \n",
       "Times Makes Two                            0.3260  \n",
       "Riding the Blues                           0.0904  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "playlist_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": false
   },
   "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>id</th>\n",
       "      <th>acousticness</th>\n",
       "      <th>danceability</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>energy</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>key</th>\n",
       "      <th>liveness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>mode</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>tempo</th>\n",
       "      <th>valence</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>The Bluest Blues (feat. George Harrison)</th>\n",
       "      <td>3d5e9kJDkgc7Jum8d8xtrl</td>\n",
       "      <td>0.5180</td>\n",
       "      <td>0.404</td>\n",
       "      <td>447240</td>\n",
       "      <td>0.331</td>\n",
       "      <td>0.353000</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0786</td>\n",
       "      <td>-11.487</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0245</td>\n",
       "      <td>73.904</td>\n",
       "      <td>0.2470</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bird of Paradise</th>\n",
       "      <td>0X7RPlCMbifRgyg1RpsR0K</td>\n",
       "      <td>0.5960</td>\n",
       "      <td>0.471</td>\n",
       "      <td>281467</td>\n",
       "      <td>0.437</td>\n",
       "      <td>0.641000</td>\n",
       "      <td>1</td>\n",
       "      <td>0.2780</td>\n",
       "      <td>-11.396</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0370</td>\n",
       "      <td>144.333</td>\n",
       "      <td>0.3370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The Thrill Is Gone</th>\n",
       "      <td>1xW9HHRQe8MMGF8SBNgkBY</td>\n",
       "      <td>0.0934</td>\n",
       "      <td>0.692</td>\n",
       "      <td>301133</td>\n",
       "      <td>0.493</td>\n",
       "      <td>0.001660</td>\n",
       "      <td>9</td>\n",
       "      <td>0.1280</td>\n",
       "      <td>-9.662</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0325</td>\n",
       "      <td>86.644</td>\n",
       "      <td>0.2800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Times Makes Two</th>\n",
       "      <td>5LJFs3O1BFipvKh2wZHNMs</td>\n",
       "      <td>0.6430</td>\n",
       "      <td>0.641</td>\n",
       "      <td>325987</td>\n",
       "      <td>0.537</td>\n",
       "      <td>0.000003</td>\n",
       "      <td>10</td>\n",
       "      <td>0.0528</td>\n",
       "      <td>-6.921</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0342</td>\n",
       "      <td>112.085</td>\n",
       "      <td>0.3260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Riding the Blues</th>\n",
       "      <td>6r8w9NhDqcMY3l8SicsaDt</td>\n",
       "      <td>0.5050</td>\n",
       "      <td>0.438</td>\n",
       "      <td>226987</td>\n",
       "      <td>0.432</td>\n",
       "      <td>0.005210</td>\n",
       "      <td>4</td>\n",
       "      <td>0.1440</td>\n",
       "      <td>-7.931</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0318</td>\n",
       "      <td>127.104</td>\n",
       "      <td>0.0904</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                              id  \\\n",
       "The Bluest Blues (feat. George Harrison)  3d5e9kJDkgc7Jum8d8xtrl   \n",
       "Bird of Paradise                          0X7RPlCMbifRgyg1RpsR0K   \n",
       "The Thrill Is Gone                        1xW9HHRQe8MMGF8SBNgkBY   \n",
       "Times Makes Two                           5LJFs3O1BFipvKh2wZHNMs   \n",
       "Riding the Blues                          6r8w9NhDqcMY3l8SicsaDt   \n",
       "\n",
       "                                          acousticness  danceability  \\\n",
       "The Bluest Blues (feat. George Harrison)        0.5180         0.404   \n",
       "Bird of Paradise                                0.5960         0.471   \n",
       "The Thrill Is Gone                              0.0934         0.692   \n",
       "Times Makes Two                                 0.6430         0.641   \n",
       "Riding the Blues                                0.5050         0.438   \n",
       "\n",
       "                                          duration_ms  energy  \\\n",
       "The Bluest Blues (feat. George Harrison)       447240   0.331   \n",
       "Bird of Paradise                               281467   0.437   \n",
       "The Thrill Is Gone                             301133   0.493   \n",
       "Times Makes Two                                325987   0.537   \n",
       "Riding the Blues                               226987   0.432   \n",
       "\n",
       "                                          instrumentalness  key  liveness  \\\n",
       "The Bluest Blues (feat. George Harrison)          0.353000    9    0.0786   \n",
       "Bird of Paradise                                  0.641000    1    0.2780   \n",
       "The Thrill Is Gone                                0.001660    9    0.1280   \n",
       "Times Makes Two                                   0.000003   10    0.0528   \n",
       "Riding the Blues                                  0.005210    4    0.1440   \n",
       "\n",
       "                                          loudness  mode  speechiness  \\\n",
       "The Bluest Blues (feat. George Harrison)   -11.487     0       0.0245   \n",
       "Bird of Paradise                           -11.396     1       0.0370   \n",
       "The Thrill Is Gone                          -9.662     0       0.0325   \n",
       "Times Makes Two                             -6.921     0       0.0342   \n",
       "Riding the Blues                            -7.931     0       0.0318   \n",
       "\n",
       "                                            tempo  valence  \n",
       "The Bluest Blues (feat. George Harrison)   73.904   0.2470  \n",
       "Bird of Paradise                          144.333   0.3370  \n",
       "The Thrill Is Gone                         86.644   0.2800  \n",
       "Times Makes Two                           112.085   0.3260  \n",
       "Riding the Blues                          127.104   0.0904  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "playlist_df=playlist_df[[\"id\", \"acousticness\", \"danceability\", \"duration_ms\", \n",
    "                         \"energy\", \"instrumentalness\",  \"key\", \"liveness\",\n",
    "                         \"loudness\", \"mode\", \"speechiness\", \"tempo\", \"valence\"]]\n",
    "playlist_df.head()"
   ]
  },
  {
   "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>id</th>\n",
       "      <th>acousticness</th>\n",
       "      <th>danceability</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>energy</th>\n",
       "      <th>instrumentalness</th>\n",
       "      <th>key</th>\n",
       "      <th>liveness</th>\n",
       "      <th>loudness</th>\n",
       "      <th>mode</th>\n",
       "      <th>speechiness</th>\n",
       "      <th>tempo</th>\n",
       "      <th>valence</th>\n",
       "      <th>ratings</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>The Bluest Blues (feat. George Harrison)</th>\n",
       "      <td>3d5e9kJDkgc7Jum8d8xtrl</td>\n",
       "      <td>0.5180</td>\n",
       "      <td>0.404</td>\n",
       "      <td>447240</td>\n",
       "      <td>0.331</td>\n",
       "      <td>0.353000</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0786</td>\n",
       "      <td>-11.487</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0245</td>\n",
       "      <td>73.904</td>\n",
       "      <td>0.2470</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bird of Paradise</th>\n",
       "      <td>0X7RPlCMbifRgyg1RpsR0K</td>\n",
       "      <td>0.5960</td>\n",
       "      <td>0.471</td>\n",
       "      <td>281467</td>\n",
       "      <td>0.437</td>\n",
       "      <td>0.641000</td>\n",
       "      <td>1</td>\n",
       "      <td>0.2780</td>\n",
       "      <td>-11.396</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0370</td>\n",
       "      <td>144.333</td>\n",
       "      <td>0.3370</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The Thrill Is Gone</th>\n",
       "      <td>1xW9HHRQe8MMGF8SBNgkBY</td>\n",
       "      <td>0.0934</td>\n",
       "      <td>0.692</td>\n",
       "      <td>301133</td>\n",
       "      <td>0.493</td>\n",
       "      <td>0.001660</td>\n",
       "      <td>9</td>\n",
       "      <td>0.1280</td>\n",
       "      <td>-9.662</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0325</td>\n",
       "      <td>86.644</td>\n",
       "      <td>0.2800</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Times Makes Two</th>\n",
       "      <td>5LJFs3O1BFipvKh2wZHNMs</td>\n",
       "      <td>0.6430</td>\n",
       "      <td>0.641</td>\n",
       "      <td>325987</td>\n",
       "      <td>0.537</td>\n",
       "      <td>0.000003</td>\n",
       "      <td>10</td>\n",
       "      <td>0.0528</td>\n",
       "      <td>-6.921</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0342</td>\n",
       "      <td>112.085</td>\n",
       "      <td>0.3260</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Riding the Blues</th>\n",
       "      <td>6r8w9NhDqcMY3l8SicsaDt</td>\n",
       "      <td>0.5050</td>\n",
       "      <td>0.438</td>\n",
       "      <td>226987</td>\n",
       "      <td>0.432</td>\n",
       "      <td>0.005210</td>\n",
       "      <td>4</td>\n",
       "      <td>0.1440</td>\n",
       "      <td>-7.931</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0318</td>\n",
       "      <td>127.104</td>\n",
       "      <td>0.0904</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                              id  \\\n",
       "The Bluest Blues (feat. George Harrison)  3d5e9kJDkgc7Jum8d8xtrl   \n",
       "Bird of Paradise                          0X7RPlCMbifRgyg1RpsR0K   \n",
       "The Thrill Is Gone                        1xW9HHRQe8MMGF8SBNgkBY   \n",
       "Times Makes Two                           5LJFs3O1BFipvKh2wZHNMs   \n",
       "Riding the Blues                          6r8w9NhDqcMY3l8SicsaDt   \n",
       "\n",
       "                                          acousticness  danceability  \\\n",
       "The Bluest Blues (feat. George Harrison)        0.5180         0.404   \n",
       "Bird of Paradise                                0.5960         0.471   \n",
       "The Thrill Is Gone                              0.0934         0.692   \n",
       "Times Makes Two                                 0.6430         0.641   \n",
       "Riding the Blues                                0.5050         0.438   \n",
       "\n",
       "                                          duration_ms  energy  \\\n",
       "The Bluest Blues (feat. George Harrison)       447240   0.331   \n",
       "Bird of Paradise                               281467   0.437   \n",
       "The Thrill Is Gone                             301133   0.493   \n",
       "Times Makes Two                                325987   0.537   \n",
       "Riding the Blues                               226987   0.432   \n",
       "\n",
       "                                          instrumentalness  key  liveness  \\\n",
       "The Bluest Blues (feat. George Harrison)          0.353000    9    0.0786   \n",
       "Bird of Paradise                                  0.641000    1    0.2780   \n",
       "The Thrill Is Gone                                0.001660    9    0.1280   \n",
       "Times Makes Two                                   0.000003   10    0.0528   \n",
       "Riding the Blues                                  0.005210    4    0.1440   \n",
       "\n",
       "                                          loudness  mode  speechiness  \\\n",
       "The Bluest Blues (feat. George Harrison)   -11.487     0       0.0245   \n",
       "Bird of Paradise                           -11.396     1       0.0370   \n",
       "The Thrill Is Gone                          -9.662     0       0.0325   \n",
       "Times Makes Two                             -6.921     0       0.0342   \n",
       "Riding the Blues                            -7.931     0       0.0318   \n",
       "\n",
       "                                            tempo  valence  ratings  \n",
       "The Bluest Blues (feat. George Harrison)   73.904   0.2470       10  \n",
       "Bird of Paradise                          144.333   0.3370        9  \n",
       "The Thrill Is Gone                         86.644   0.2800        9  \n",
       "Times Makes Two                           112.085   0.3260       10  \n",
       "Riding the Blues                          127.104   0.0904        8  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Change the below info for each playlist manually to get better results\n",
    "# Give ratings to your tracks with respect to their playlist relevances\n",
    "# Rate them from 1-10, give higher ratings to those tracks which you think best chracterizes your playlist\n",
    "# If you order your playlist by relevance while creating it, this step will become easier\n",
    "# So now, we will deal with a classification task\n",
    "import numpy as np\n",
    "playlist_df['ratings']=[10, 9, 9, 10, 8, 6, 8, 4, 3, 5, 7, 5, 5, 8, 8, 7, 8, 8, 10, 8, 10, 8, 4, 4, 4, 10, 10, 9, 8, 8, 4]\n",
    "playlist_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature ranking:\n",
      "1. acousticness 0.186450 \n",
      "2. danceability 0.138950 \n",
      "3. duration_ms 0.109225 \n",
      "4. energy 0.103662 \n",
      "5. instrumentalness 0.101786 \n",
      "6. key 0.069612 \n",
      "7. liveness 0.067485 \n",
      "8. loudness 0.060709 \n",
      "9. mode 0.053132 \n",
      "10. speechiness 0.048724 \n",
      "11. tempo 0.043043 \n",
      "12. valence 0.017224 \n"
     ]
    }
   ],
   "source": [
    "# Analyze feature importances\n",
    "from sklearn.ensemble.forest import RandomForestRegressor, RandomForestClassifier\n",
    "\n",
    "X_train = playlist_df.drop(['id', 'ratings'], axis=1)\n",
    "y_train = playlist_df['ratings']\n",
    "forest = RandomForestClassifier(random_state=42, max_depth=5, max_features=12) # Set by GridSearchCV below\n",
    "forest.fit(X_train, y_train)\n",
    "importances = forest.feature_importances_\n",
    "indices = np.argsort(importances)[::-1]\n",
    "\n",
    "# Print the feature rankings\n",
    "print(\"Feature ranking:\")\n",
    "  \n",
    "for f in range(len(importances))\n",
    "    print(\"%d. %s %f \" % (f + 1, \n",
    "            X_train.columns[f], \n",
    "            importances[indices[f]]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2c38b76bba8>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 425,
       "width": 612
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Apply pca to the scaled train set first\n",
    "\n",
    "from sklearn import decomposition\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns; sns.set(style='white')\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "X_scaled = StandardScaler().fit_transform(X_train)\n",
    "\n",
    "pca = decomposition.PCA().fit(X_scaled)\n",
    "\n",
    "plt.figure(figsize=(10,7))\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_), color='k', lw=2)\n",
    "plt.xlabel('Number of components')\n",
    "plt.ylabel('Total explained variance')\n",
    "plt.xlim(0, 12)\n",
    "plt.yticks(np.arange(0, 1.1, 0.1))\n",
    "plt.axvline(8, c='b') # Tune this so that you obtain at least a 95% total variance explained\n",
    "plt.axhline(0.95, c='r')\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fit your dataset to the optimal pca\n",
    "pca1 = decomposition.PCA(n_components=8)\n",
    "X_pca = pca1.fit_transform(X_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_score, StratifiedKFold, GridSearchCV\n",
    "from sklearn.metrics import accuracy_score\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "# Initialize a stratified split for the validation process\n",
    "skf = StratifiedKFold(n_splits=5, shuffle=True, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:   23.5s\n",
      "[Parallel(n_jobs=-1)]: Done  90 out of  90 | elapsed:   29.5s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "             max_depth=5, max_features=4, max_leaf_nodes=None,\n",
       "             min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "             min_samples_leaf=1, min_samples_split=2,\n",
       "             min_weight_fraction_leaf=0.0, n_estimators=100, n_jobs=-1,\n",
       "             oob_score=True, random_state=42, verbose=0, warm_start=False),\n",
       " 0.3548387096774194)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Initialize the set of parameters for exhaustive search and fit \n",
    "parameters = {'max_features': [4, 7, 8], 'min_samples_leaf': [1, 3, 5], 'max_depth': [3,5]}\n",
    "rfc = RandomForestClassifier(n_estimators=100, random_state=42, \n",
    "                             n_jobs=-1, oob_score=True)\n",
    "gcv1 = GridSearchCV(rfc, parameters, n_jobs=-1, cv=skf, verbose=1)\n",
    "gcv1.fit(X_pca, y_train)\n",
    "gcv1.best_estimator_, gcv1.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# You may want to check the results of TSNE also\n",
    "from sklearn.manifold import TSNE\n",
    "tsne = TSNE(random_state=17)\n",
    "\n",
    "X_tsne = tsne.fit_transform(X_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(31, 2)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_tsne.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 4 candidates, totalling 20 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  20 out of  20 | elapsed:   21.1s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "             max_depth=1, max_features=1, max_leaf_nodes=None,\n",
       "             min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "             min_samples_leaf=1, min_samples_split=2,\n",
       "             min_weight_fraction_leaf=0.0, n_estimators=100, n_jobs=-1,\n",
       "             oob_score=True, random_state=42, verbose=0, warm_start=False),\n",
       " 0.2903225806451613)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parameters = {'max_features': [1,2], 'min_samples_leaf': [1,2], 'max_depth': [1,2]}\n",
    "gcv2 = GridSearchCV(rfc, parameters, n_jobs=-1, cv=skf, verbose=1)\n",
    "gcv2.fit(X_tsne, y_train)\n",
    "gcv2.best_estimator_, gcv2.best_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Don't expect high cross-val scores! Our dataset is way too small...\n",
    "For my case, PCA + rfc seems to perform better than worst.."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>analysis_url</th>\n",
       "      <th>danceability</th>\n",
       "      <th>duration_ms</th>\n",
       "      <th>energy</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>speechiness</th>\n",
       "      <th>tempo</th>\n",
       "      <th>time_signature</th>\n",
       "      <th>track_href</th>\n",
       "      <th>type</th>\n",
       "      <th>uri</th>\n",
       "      <th>valence</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3ZewawA5h2yP6RvhWlBdFq</th>\n",
       "      <td>0.651</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/3Zew...</td>\n",
       "      <td>0.466</td>\n",
       "      <td>405160</td>\n",
       "      <td>0.342</td>\n",
       "      <td>3ZewawA5h2yP6RvhWlBdFq</td>\n",
       "      <td>0.005820</td>\n",
       "      <td>9</td>\n",
       "      <td>0.1010</td>\n",
       "      <td>-8.166</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0372</td>\n",
       "      <td>121.252</td>\n",
       "      <td>3</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/3ZewawA5h2yP...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:3ZewawA5h2yP6RvhWlBdFq</td>\n",
       "      <td>0.118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5e0Yfg4VsZ23o4bJ8TIe7Q</th>\n",
       "      <td>0.741</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/5e0Y...</td>\n",
       "      <td>0.441</td>\n",
       "      <td>243800</td>\n",
       "      <td>0.340</td>\n",
       "      <td>5e0Yfg4VsZ23o4bJ8TIe7Q</td>\n",
       "      <td>0.000271</td>\n",
       "      <td>5</td>\n",
       "      <td>0.1240</td>\n",
       "      <td>-9.606</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0310</td>\n",
       "      <td>73.860</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/5e0Yfg4VsZ23...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:5e0Yfg4VsZ23o4bJ8TIe7Q</td>\n",
       "      <td>0.138</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4nvcK9IdGNjfcZlNd4vlE4</th>\n",
       "      <td>0.370</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/4nvc...</td>\n",
       "      <td>0.440</td>\n",
       "      <td>317640</td>\n",
       "      <td>0.508</td>\n",
       "      <td>4nvcK9IdGNjfcZlNd4vlE4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0987</td>\n",
       "      <td>-6.389</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0275</td>\n",
       "      <td>121.188</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/4nvcK9IdGNjf...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:4nvcK9IdGNjfcZlNd4vlE4</td>\n",
       "      <td>0.192</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7xPeBTWPq4xNgQa4g1qA0e</th>\n",
       "      <td>0.353</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/7xPe...</td>\n",
       "      <td>0.526</td>\n",
       "      <td>276333</td>\n",
       "      <td>0.578</td>\n",
       "      <td>7xPeBTWPq4xNgQa4g1qA0e</td>\n",
       "      <td>0.000217</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0689</td>\n",
       "      <td>-7.084</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0296</td>\n",
       "      <td>129.957</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/7xPeBTWPq4xN...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:7xPeBTWPq4xNgQa4g1qA0e</td>\n",
       "      <td>0.562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>634ENkvsEK6O8Gs4A5E6Zl</th>\n",
       "      <td>0.712</td>\n",
       "      <td>https://api.spotify.com/v1/audio-analysis/634E...</td>\n",
       "      <td>0.245</td>\n",
       "      <td>191987</td>\n",
       "      <td>0.468</td>\n",
       "      <td>634ENkvsEK6O8Gs4A5E6Zl</td>\n",
       "      <td>0.127000</td>\n",
       "      <td>2</td>\n",
       "      <td>0.4590</td>\n",
       "      <td>-9.056</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0310</td>\n",
       "      <td>132.529</td>\n",
       "      <td>4</td>\n",
       "      <td>https://api.spotify.com/v1/tracks/634ENkvsEK6O...</td>\n",
       "      <td>audio_features</td>\n",
       "      <td>spotify:track:634ENkvsEK6O8Gs4A5E6Zl</td>\n",
       "      <td>0.115</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        acousticness  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq         0.651   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q         0.741   \n",
       "4nvcK9IdGNjfcZlNd4vlE4         0.370   \n",
       "7xPeBTWPq4xNgQa4g1qA0e         0.353   \n",
       "634ENkvsEK6O8Gs4A5E6Zl         0.712   \n",
       "\n",
       "                                                             analysis_url  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq  https://api.spotify.com/v1/audio-analysis/3Zew...   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q  https://api.spotify.com/v1/audio-analysis/5e0Y...   \n",
       "4nvcK9IdGNjfcZlNd4vlE4  https://api.spotify.com/v1/audio-analysis/4nvc...   \n",
       "7xPeBTWPq4xNgQa4g1qA0e  https://api.spotify.com/v1/audio-analysis/7xPe...   \n",
       "634ENkvsEK6O8Gs4A5E6Zl  https://api.spotify.com/v1/audio-analysis/634E...   \n",
       "\n",
       "                        danceability  duration_ms  energy  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq         0.466       405160   0.342   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q         0.441       243800   0.340   \n",
       "4nvcK9IdGNjfcZlNd4vlE4         0.440       317640   0.508   \n",
       "7xPeBTWPq4xNgQa4g1qA0e         0.526       276333   0.578   \n",
       "634ENkvsEK6O8Gs4A5E6Zl         0.245       191987   0.468   \n",
       "\n",
       "                                            id  instrumentalness  key  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq  3ZewawA5h2yP6RvhWlBdFq          0.005820    9   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q  5e0Yfg4VsZ23o4bJ8TIe7Q          0.000271    5   \n",
       "4nvcK9IdGNjfcZlNd4vlE4  4nvcK9IdGNjfcZlNd4vlE4          0.000000    0   \n",
       "7xPeBTWPq4xNgQa4g1qA0e  7xPeBTWPq4xNgQa4g1qA0e          0.000217    1   \n",
       "634ENkvsEK6O8Gs4A5E6Zl  634ENkvsEK6O8Gs4A5E6Zl          0.127000    2   \n",
       "\n",
       "                        liveness  loudness  mode  speechiness    tempo  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq    0.1010    -8.166     0       0.0372  121.252   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q    0.1240    -9.606     1       0.0310   73.860   \n",
       "4nvcK9IdGNjfcZlNd4vlE4    0.0987    -6.389     1       0.0275  121.188   \n",
       "7xPeBTWPq4xNgQa4g1qA0e    0.0689    -7.084     1       0.0296  129.957   \n",
       "634ENkvsEK6O8Gs4A5E6Zl    0.4590    -9.056     1       0.0310  132.529   \n",
       "\n",
       "                        time_signature  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq               3   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q               4   \n",
       "4nvcK9IdGNjfcZlNd4vlE4               4   \n",
       "7xPeBTWPq4xNgQa4g1qA0e               4   \n",
       "634ENkvsEK6O8Gs4A5E6Zl               4   \n",
       "\n",
       "                                                               track_href  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq  https://api.spotify.com/v1/tracks/3ZewawA5h2yP...   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q  https://api.spotify.com/v1/tracks/5e0Yfg4VsZ23...   \n",
       "4nvcK9IdGNjfcZlNd4vlE4  https://api.spotify.com/v1/tracks/4nvcK9IdGNjf...   \n",
       "7xPeBTWPq4xNgQa4g1qA0e  https://api.spotify.com/v1/tracks/7xPeBTWPq4xN...   \n",
       "634ENkvsEK6O8Gs4A5E6Zl  https://api.spotify.com/v1/tracks/634ENkvsEK6O...   \n",
       "\n",
       "                                  type                                   uri  \\\n",
       "3ZewawA5h2yP6RvhWlBdFq  audio_features  spotify:track:3ZewawA5h2yP6RvhWlBdFq   \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q  audio_features  spotify:track:5e0Yfg4VsZ23o4bJ8TIe7Q   \n",
       "4nvcK9IdGNjfcZlNd4vlE4  audio_features  spotify:track:4nvcK9IdGNjfcZlNd4vlE4   \n",
       "7xPeBTWPq4xNgQa4g1qA0e  audio_features  spotify:track:7xPeBTWPq4xNgQa4g1qA0e   \n",
       "634ENkvsEK6O8Gs4A5E6Zl  audio_features  spotify:track:634ENkvsEK6O8Gs4A5E6Zl   \n",
       "\n",
       "                        valence  \n",
       "3ZewawA5h2yP6RvhWlBdFq    0.118  \n",
       "5e0Yfg4VsZ23o4bJ8TIe7Q    0.138  \n",
       "4nvcK9IdGNjfcZlNd4vlE4    0.192  \n",
       "7xPeBTWPq4xNgQa4g1qA0e    0.562  \n",
       "634ENkvsEK6O8Gs4A5E6Zl    0.115  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Now build your test set;\n",
    "# Generate a new dataframe for recommended tracks\n",
    "# Set recommendation limit as half the Playlist Length per track, you may change this as you like\n",
    "# Check documentation for  recommendations; https://beta.developer.spotify.com/documentation/web-api/reference/browse/get-recommendations/\n",
    "\n",
    "rec_tracks = []\n",
    "for i in playlist_df['id'].values.tolist():\n",
    "    rec_tracks += sp.recommendations(seed_tracks=[i], limit=int(len(playlist_df)/2))['tracks'];\n",
    "\n",
    "rec_track_ids = []\n",
    "rec_track_names = []\n",
    "for i in rec_tracks:\n",
    "    rec_track_ids.append(i['id'])\n",
    "    rec_track_names.append(i['name'])\n",
    "\n",
    "rec_features = []\n",
    "for i in range(0,len(rec_track_ids)):\n",
    "    rec_audio_features = sp.audio_features(rec_track_ids[i])\n",
    "    for track in rec_audio_features:\n",
    "        rec_features.append(track)\n",
    "        \n",
    "rec_playlist_df = pd.DataFrame(rec_features, index = rec_track_ids)\n",
    "rec_playlist_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "rec_playlist_df=rec_playlist_df[[\"acousticness\", \"danceability\", \"duration_ms\", \n",
    "                         \"energy\", \"instrumentalness\",  \"key\", \"liveness\",\n",
    "                         \"loudness\", \"mode\", \"speechiness\", \"tempo\", \"valence\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Make predictions\n",
    "gcv1.best_estimator_.fit(X_pca, y_train)\n",
    "rec_playlist_df_scaled = StandardScaler().fit_transform(rec_playlist_df)\n",
    "rec_playlist_df_pca = pca1.transform(rec_playlist_df_scaled)\n",
    "y_pred_class = gcv1.best_estimator_.predict(rec_playlist_df_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "rec_playlist_df['ratings']=y_pred_class\n",
    "rec_playlist_df = rec_playlist_df.sort_values('ratings', ascending = False)\n",
    "rec_playlist_df = rec_playlist_df.reset_index()\n",
    "\n",
    "# Pick the top ranking tracks to add your new playlist 9, 10 will work\n",
    "recs_to_add = rec_playlist_df[rec_playlist_df['ratings']>=9]['index'].values.tolist()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(465, (465, 15), 78)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check what is about to happen :)\n",
    "len(rec_tracks), rec_playlist_df.shape, len(recs_to_add)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a new playlist for tracks to add - you may also add these tracks to your source playlist and proceed\n",
    "playlist_recs = sp.user_playlist_create(username, \n",
    "                                        name='More Recommended Songs for Playlist {}'.format(sourcePlaylist['name']))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add tracks to the new playlist\n",
    "sp.user_playlist_add_tracks(username, playlist_recs['id'], recs_to_add);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This performed really well for my case, and I am pretty satisfied with the resulting playlist -more than the default recommendations"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}