What we tweet about when we tweet about taxes: A topic modelling approach

•Using an unsupervised machine learning approach, we identified several content topics within over a million tax related tweets.•We classified people's conversations on Twitter in the following groups: 1. Opinions about tax Politics, 2. Enforcement, 3. Information & Service, and 4. Emotions.•The most frequently occurring tweets are about giving or seeking information and services, with a steady increase during the last decade.•The findings favor the idea of policies promoting the “service and client” approach, which several tax offices are adopting.