Datasets

Datasets for the tutorial

Magnatagatune - We are using data from the magnatagatune project.  In particular,  the echonest xml data, and the annotation (tags) set.  It is also worthwhile to download all of the associated magnatune music via bittorrent.  The songs are 30 second mono clips, and total just under 3GB.   If you want the scripts to work automatically, you will need to place all of the data in the same folder (i.e. you should have folders "mp3", "mp3_echonest_xml", and the file "annotations_final.csv" in the same folder).

ArtistSimilarityDat - We are also using Echo Nest artist similarity data.  This data was collected during the spring of 2009. It contains similar artist, familiarity and hotness data for approximately 70,000 artists.   The data is available here: ArtistSimilarityDat.tar.gz