Select a Bandcamp library/label for clustering and classification according to Beatport's taxonomy composed of 32 subgenres (as of Jan 2021).
Select the number of subgenre clusters into which the libraries/labels will be distributed.
The default number of clusters is 4, but you can also try out other values. Use 1 cluster for analysing the overall sound of the label/library.
Use "Auto" for automatic cluster number recommendation (note: this option will take more time to process).
Note: the K-means clustering algorithm used in this project may give different results every time.
The classification relies on audio features extracted individually from each label's tracks.
You will receive the results in a PDF file with links to up to three representative Bandcamp tracks (if possible, from different artists) in each group.
The document will also show the tags (folksonomies) added by the artists/labels to the Bandcamp pages, enabling comparison with Beatport's subgenre taxonomy.
You can add additional libraries to the project on the Admin interface.
For more info please refer to the Github site of the project.