Data was anonymised prior to cleaning, through random re-assignment of unique IDs and removal of potentially identifying variables. Aside from the anonymisation changes, the raw data is provided as well as the code used to clean data to produce the aggregate data used for analysis.