Pandora is an interesting alternative to last.fm, which offers music recommendations based on what you’re listening to. While last.fm is one giant data mining application based only (I think) on the actual playlists of their users, Pandora uses data from the Music Genome Project. Actual people listened to thousands of bands and songs and tried to identify characteristic attributes of their music. While Pandora is probably a lot less “Web 2.0″ than last.fm and (potentially) not that unlimited in scope, it might work better than the statistical approach taken by last.fm. Some of my favorite bands were classified as follows:
- Nine Inch Nails: electronica influences, minory key tonality, use of techno synths, a subtle use of vocal harmony, extensive vamping, a vocal-centric aesthetic and a clear focus on recording studio production. This leads to recommendations like Gary Numan, Kovenant or Type’O'Negative.
- Tool: Hard rock roots, a subtle use of vocal harmony, a vocal-centric aesthetic, minor key tonality.
- Beastie Boys: East coast rap influences, clean lyrics, bumping’ kick sound, use of background scratching, synth tweaking, a dry recording sound, trippy soundscapes, acoustic drum samples, and the subtle use of turntables.
- Placebo: Composed melodic style, a vocal-centric aesthetic, major key tonality, basic rock song structures.
The quality of Pandora’s recommendations oscillates between “interesting” and “err… no”, but it’s definitely worth a try for finding new music.