Working with lastfm data to cluster like artists together based on genre and user plays.
http://www.dtic.upf.edu/~ocelma/MusicRecommendationDataset/
usersha1-profile.tsv userhash gender age country join date
usersha1-artmbid-artname-plays.tsv userhash artist_musicbrainz_id artistname plays
359349 unique user hashes 160168 unique artist hashes 17559530 rows
>>> 17559530/float(359347*(186642+107373))
0.00016619937111663383
kaggle_visible_evaluation.txt
110000 unique user hashes 163206 unique song hashes 1450934 rows
>>> 1450934/float(n*m)
8.082000104719857e-05