So what do social media data (not) tell us about the historical event? Reflections on a corpus of Twitter data in the context of the Greek debt crisis of the 2010s