The project analysed the performance of community detection algorithms on the Twitter social network operating on a graph compressed using minhash signatures. The data supplied gives minhash signatures of roughly 16,000 Twitter users who have been classified into 16 categories. It is described in https://arxiv.org/abs/1601.03958 and together with code at https://github.com/melifluos/LSH-community-detection allows the results within to be replicated.