post-template-default single single-post postid-23436 single-format-standard

Twitter ‘Hate Map’ Shows where Racist, Homophobic and Offensive Tweets Originate

Alex Constantine - May 20, 2013

Students at Humboldt State University in California individually reviewed 150,000 geocoded tweets containing racist, homophobic, or otherwise offensive terms to build a “hate map” indicating where people in the U.S. are  most bigoted.

Or, at least, where they’re the most open about displaying their antisocial views.

apdh 300x167 - Twitter ‘Hate Map’ Shows where Racist, Homophobic and Offensive Tweets OriginateThe picture doesn’t look good for the Eastern states, although admittedly the  bulk of the population is there as well. Areas in Virginia, North Carolina, Texas, and Alabama show up bright red on the map, as do areas in more central  states Indiana, Iowa, and Minnesota.

The map is part of a larger project, called the Geography of Hate, by  Humboldt State professor Dr. Monica Stephens. The data that forms the map comes  from an analysis of every tweet posted between June 2012 and April 2013  that contained at least one of 10 designated “hate words,” including dyke, fag,  chink, gook, wetback, and cripple.

But while the original list of tweets was generated by a machine, every  single one of the 150,000 tweets containing one of the target words was  individually examined by undergraduate students. As the project description  states:

Because algorithmic sentiment analysis would automatically classify any tweet  containing “hate words” as “negative,” this project relied upon the HSU students  to read the entirety of tweet and classify it as positive, neutral or negative  based on a predefined rubric. Only those tweets that were identified by human  readers as negative were used in this analysis.

To protect the identity of potentially racist, homophobic, or otherwise  bigoted Twitter users, the tweets were aggregated up to the county level, and  counties with high levels of hate speech were colored red on the map. Areas with  moderate levels — though still higher than the national average — are varying  shades of blue, and unshaded areas were below the national average.

Smaller towns seem to have a higher incidence of hate speech — in Virginia,  for example, Palmyra is more hateful on Twitter than Richmond. And in Louisiana,  New Orleans and Baton Rouge are less hateful than smaller towns nearby.

Image credits: Geography of Hate

Read more at homophobic-and-offensive-tweets-originate/#YkQeATjVSD980Y36.99

Leave a Reply

Your email address will not be published. Required fields are marked *