Note: data files are too large to upload directly for github Data Source: Spotify Million Playlist Dataset
Goal: Given a base playlist, can I figure out "similar" songs to add to playlist?
Testing: Take a playlist and remove 30% of the songs. Then see how many of the removed songs I can accurately add to the playlist.
Layer 1: Find "similar" playlist based on song intersection with base playlist
Layer 2: Find "similar" songs to add using spotify api
Weights: Use # of followers to create weighting scheme for playlist importance