Mining the MySpace social graph
One really cool source of music data is this live RDF representation of Myspace users.
This RDF representation gives you a FOAF-style view of a myspace user.
If the user happens to be a music artist, the corresponding tracks for
the artist included in the RDF. This gives you a nice way of
extracting the artist social graph at MySpace - which can be used to do
all sorts of things from artist similarity to playlist generation. The
code is open sourced.Update Yves and Kurt are responsible for the RDF translation service with some recent input from Mischa Tuffield of Garlik (to help to refine the service and the supporting ontology). Ben Fields did loads of work gathering the data set and extracting audio features.
Keep an eye on these guys, they are going places!
The other dbtune services also provide quite a lot of music-related RDF data (over 15 billion triples). All the datasets provide links to other dataset (eg. Musicbrainz artists' URIs are linked to the MySpace artists' URIs).
Quite a lot of linked data to play with :-)
Posted by yves on September 17, 2008 at 04:08 PM EDT #
Actually Yves Raimond and myself are responsible for the RDF translation service (so if it breaks holler;) Also some recent input from Mischa Tuffield of Garlik has helped to refine the service and the supporting ontology. Of course Ben did loads of work gathering the data set and extracting audio features. Thanks for the mention Paul!
Posted by Kurt Jacobson on September 17, 2008 at 04:28 PM EDT #