According to an example, synthetic time series data generation may include receiving time series data for a plurality of users, and applying dimensionality reduction to transform the time series data from a high dimensional space n of the time series data to a low dimensional space m, where m