The repository contains the raw data on which the analysis is based, the code used to reproduce the results, and the trained topic models.