The part I'm most curious about is how Floating Sheep was looking for tweets "with racist terms that also reference the election". It just seems like 395 geocod

The part I'm most curious about is how Floating Sheep was looking for tweets "with racist terms that also reference the election". It just seems like 395 geocoded racist tweets is low. If they were just running tweets against a whitelist of things they thought a racist tweet might contain there could have been a lot of tweets falling through the cracks. Or, hey, maybe not. The comments on Mashable and the original Floating Sheep article sure did get ugly quick though. EDIT: A little bit of digging in the comments turned up a follow-up post by Floating Sheep: http://www.floatingsheep.org/2012/11/faq-mapping-racist-tweets-in-response.html http://www.floatingsheep.org/2012/11/faq-mapping-racist-twee...