77percent™

yea, pandora’s great for discovering
and also just for quickly putting something on that you know you’ll probably like but also probably won’t be able to see coming … it’s just that it can be very limited by having such little information to go by, i’m very impressed with how accurately it can describe an artist, but firstly it’s dependent on the knowledge of musical professionals behind the music genome project which is great and all but limits it’s ability to find new artists and also potentially introduces bias (i.e it’s not as if they really listen to (for themselves) everything that they analyze (i’m not suggesting that they make up data, but rather that the music they analyze might not be the music they listen to personally)) .. i don’t really know where i’m going with this but i like how musical similarity/etc is determined by what people are listening. it seems awkward having someone analyze music that they don’t necessarily enjoy listening to.

i really apreciate last.fm’s collaborative-ness , in that first of all it has all of my music listening history to look at in comparison to a whole lot of other people’s listening habits to analyze .. i really appreciate collaborative filtering
i kind of like the idea of ‘
what people who like what you like tend to like‘ , because it’s not just a matter of what the most popular music out there is [i.e determined by radio stations (payola) or music professionals’ analysis) but rather the music that is being discovered/listenedto by people who have similar tendancies to you

i also really like having records of all the songs i listen to , it’s a very cool application of stats :it is enormously difficult to find a cool application for statistical analytics)

i like to think that music can really define a lot about a person and it’s nice to meet people through music (a lot of people have met (myself included) people they really get along with
at concerts /etc)