this disease of mapping tweets and then inferring all sorts of nonsense from it has to stop. you start with a extraordinarily low quality dataset and then mash it with inaccurate geolocation, and you get j-school drivel.

• Uncategorized