The Geography of Hate Speech on Twitter
Dr. Monica Stephens, professor at Humboldt State University in California, worked with undergraduate researchers to create The Geography of Hate Map. The map geographically tags and plots homophobic and racist statements tweeted all over America from June 2012 - April 2013.
In Stephens’ introduction to the map, she explains that HSU collected the data with DOLLY (Data On Local Life and You), a University of Kentucky project that maps social media geography for research.
The Geography of Hate Map suggests that out of 150,000 mapped tweets, most haters reign from the Midwest to the East Coast. Is this accurate? Sort of.
Stephens herself notes, “Even when normalized, many of the slurs included in our analysis display little meaningful spatial distribution,” and as she later tweeted, “in the east coast the counties are smaller so if a word is used in adjacent counties it appears as a hotspot,” which accounts for some of the East Coast / West Coast disparity.
What about hate words that are used in a joking way? As Chris Rock points out in his stand-up: ”It’s not always the word [that’s offensive], it’s the context in which the word is said.” To account for such varying intent, the researchers read each “hate-tweet” individually to determine a tweet’s sentiment as positive, negative, or neutral — and only negative tweets are shown on the map.
Though the study accurately depicts the hate of those Tweeters that managed to make it into the study, the map isn’t a perfect depiction of Twitter hate in the US. As Matt Peckham notes: people who haven’t enabled geotagging aren’t included in the study, meaning there could be more hateful tweets out there that haven’t been plotted. Also, more hate words exist than those Stephens chose to incorporate; when those other hate words aren’t counted, results are skewed.
FJP: When social media becomes social meanie-a… - Krissy
Image: Screenshot of The Geography of Hate Map