Datafiles for the MovieLens dataset for benchmarking purposes.

These are slightly modified versions of the originals so as to easy the import process. The original MovieLens datasets are over here: http://www.grouplens.org/datasets/movielens

Many of the benchmarks here are for the bcolz package, and are strongly based on Greg Redas's previous work: http://www.gregreda.com/2013/10/26/using-pandas-on-the-movielens-dataset/

Here you can find the IPython notebook that was presented in EuroPython 2014:

http://nbviewer.ipython.org/github/Blosc/movielens-bench/blob/master/querying-ep14.ipynb