4 ms·
I make extensive use of the Spotify algorithm to discover the music I want to listen to based on my own idea of what it should sound like. For example if I want
by wdutch 4y ago
I make extensive use of the Spotify algorithm to discover the music I want to listen to based on my own idea of what it should sound like. For example if I want some music with a house beat but folky feel to it, then I'll create a playlist and 'seed' it with some house music and some folk music; about 10 tracks total. Then I'll scroll to the very bottom of the playlist and go through the Recommended Songs. Spotify is often pretty good at finding an 'average' of the different songs added to the playlist. I add the songs that fit my expectations to the playlist which further refines the recommendations. After I've added about 10 songs I'll remove the songs I used to seed the recommendations.
Sometimes this method doesn't work because maybe the 'average' you're looking for doesn't exist or Spotify hasn't connected the dots. Sometimes a track in the 'seed' can dominate the recommendations so try removing tracks that seem to skew the results or adding more from the opposite genre.
- wartijn_ 4y agoThat sounds like a fun way to discover new music. I'm definitely going to try that.
- sshine 4y agoI also make use of “song radio”, i.e. auto-generate a playlist based on a single song. I might call your method “playlist radio”, and I don’t use it as often. However, judging by how many songs on these recommender radios that I’ve already liked, I think that recommendations are somehow always based on a vector.