Spotify Features - Tinkering tutorial/ walkthrough

File descriptions

• \data contains json song data
• get_data.py allows for easy attraction of playlist data using Spotify for developers account (see below for more details)
• tinker.ipynb takes json feature data from a playlist and converts to a pandas dataframe, draws feature histograms of said data

Directions for getting song features - using Spotify for Developers Account

• you need a spotify CLIENT_ID & CLIENT_SECRET for this
• make a credentials.py file and input details, sample format below, you really just need the CLIENT_ID and CLIENT_SECRET
• also input a username for the user you want to analyze

PORT_NUMBER = 8080
CLIENT_ID = 'YOUR_CLIENT_ID_HERE'
CLIENT_SECRET = 'YOUR_CLIENT_SECRET_HERE'
REDIRECT_URI = 'http://localhost:8080'
SCOPE = 'user-library-read'
CACHE = '.spotipyoauthcache'

USERNAME = "YOUR_USERNAME_HERE"

• Create a Spotify for Developers account
• Create an app to get CLIENT_ID & CLIENT_SECRET (this we saved in the credentials.py file)
• to get USERNAME go to Spotify Web Console and go to a user, click on the three dots button > Share > Copy Spotify URI
• run get_data.py to get user data into dataframes
• main(username, client_ID, client_secret) returns two dataframes: one df with tracks data, one df with features of said tracks data
• main2(username, client_ID, client_secret, print_to_file = False): returns a text file or prints out all the tracks in the public playlists of a user in a semi-nicely formatted manner
• get_playlist_content_from_id(playlistID, sp, playlistName = "") just get the playlist track content from one playlist
• get_all_playlist_content(username, sp) gets all the tracks in the public playlists of a user, returns a df
• get_user_playlists(username, sp) gets all playlists of a user, returns a df
• get_all_track_features(tracks_df, sp) gets all track features of a dataframe of tracks, returns a track_df
• get_track_features(trackID, sp) gets audio features for a single track, returns a features_df

Directions for getting features from a playlist - without having a Spotify for Developers Account

• refer to tinker.ipynb
• Spotify Web Console > Playlists > Get a Playlist's Tracks
• Share > Copy Spotify URI. From the URI grab the playlist ID (spotify:user:username:playlist:playlist_id)
• user_id and playlist_id > items(track(album(name, href), href, id, name)) in fields.
• under OAuth Token click TRY IT
• song data in songs.json
• Get Audio Features, navigate to Tracks > Get Audio Features for Several Tracks and enter the song ids
• copy and paste feature data in features.json

References:

• Spotipy Docs
• github code reference for spotipy usage in get_data.py

To do:

• Code cleanup:
  • add a package requirements file