Spotify’s team of music data scientists are using an old school method – user surveys – to fine tune their advanced music recommendation algorithm that already seems to know a lot more about what I really want to listen to than I do.
Spotify is testing a short music survey designed to fine tune individualized music recommendations. After connecting their Spotify accounts, users are asked to rate 10 recently played track on a scale ranging from NEVER to LOVE. Rating additional 10 song batches will presumably further fine tune future recommendations.
Any Spotify user can take the survey here.
— Paul Lamere (@plamere) July 17, 2017