The Geography of Hate project mapped 150,000 tweets that used racist or homophobic terms.

This map shows tweets with the word "fag."

The maps were put together by Monica Stephens an assistant professor at Humboldt State.

This map shows tweets with the word "dyke."

To make sense of the 150,000 tweets, three of Stephens’ students (Amelia Egle, Miles Ross and Matthew Eiben) first had to find only the negative ones.

This map shows tweets with the word "homo."

They didn’t want phrases celebrating “dykes on bikes #SFPride.”

This map shows tweets with the word "queer."

Stephens says the job took about 150 hours in total, but was worth it because algorithmic analysis couldn’t have achieved the same thing.

This map shows tweets with the word "nigger."

This map shows tweets with the word "chink."

This map shows tweets with the word "gook."

This map shows tweets with the word "spick."

This map shows tweets with the word "wetback."

This map shows tweets with the word "cripple."

2013-05-15

Mapping The Most Hate-Filled Places In America

The maps from Geography of Hate look at where across the country people are most likely to be tweeting something deeply hateful.

If you’ve wondered where the truly hate-filled people in America live, take a look at these maps. They’re based on an analysis of 150,000 geo-located tweets from June last year to this April, and show where people are using Twitter to voice off homophobic and racist slurs.

It’s called the Geography of Hate. You can see, for example, that parts of Indiana and Iowa have the greatest numbers of people using the word "nigger" negatively, and that there are slightly more people using the words "homo" or "fag" in Omaha and Green Bay (though many places use them).

A map showing homophobic tweets with the word "dyke" (but not "dykes on bikes").

You can search for "dyke," "fag," "homo," and "queer" in the homophobic category, "chink," "gook," "nigger," "wetback," and "spick" in the racist column, and "cripple" under disability. The maps show relative use: where people tweeted the offending terms most often. The worst are in red, the moderately bad are in blue, with the least bad are unshaded.

The maps were put together by Monica Stephens, an assistant professor at Humboldt State. "As all the data is normalized based on how much content there is on
Twitter, the map primarily shows areas where the proportion of content is above average," she says. "As you’ll note from the maps, racist and homophobic sentiment exists around the country and not just in red states."

The map showing racist tweets with the word "nigger."

Though the maps make use of new technology like Twitter and Google’s Map API, they show some limitations, too. To make sense of the 150,000 tweets, three of Stephens’ students (Amelia Egle, Miles Ross and Matthew Eiben) first had to find only the negative ones (they didn’t want phrases celebrating “dykes on bikes #SFPride”). Stephens says the job took about 150 hours in total, but was worth it because algorithmic analysis couldn’t have achieved the same thing.

"Computers are poor judges of sarcasm, and Twitter is often used in sarcastic ways, or used to quote somebody else. Most sentiment analyzers would put all words that include these terms in the negative category.

Add New Comment

16 Comments

  • Joe

    I grew up on the East Coast for the first half of my life. Then moved to the west coast. I remember saying things on the east coast like "I'm gonna kill you!" then we would shove each other around at school as kids and it was no big deal. I said that in WA state when I first moved there and other kids were scared of me. I got reported to the principle and nearly got suspended. I thought, "what's the big deal?"  Like I'm gonna do it?

    There was a definite difference as to the weight of my words even then in the early 1990's before all the shootings and stuff.

    I also experienced a ton of racism directed at me growing up. But then I had friends of that same race who would take up for me. I found the east coast to me a much more direct and violent culture than the west coast.

  • jamezb

    In regards to the accuracy of your data, I noticed that on the chart for "dyke" there are areas along the Mississippi river where the river "dikes" have caused frequent flooding are labeled as being hotspots for hate speech, when in reality that term is rarely used for lesbian in those areas. I also see a hot spot in my state for the "n" word, where not only are there almost no residents, there is no wireless or internet service period! Clearly your analysis suffers from some serious flaws.

  • mtwzzyzx

    This whole mapping project is ridiculous- Mapping the tweets of the N-word to measure hate when there is a whole community of people who use the word in a friendly way, and frequently (the hip-hop/rap world)? Utterly meaningless.

  • Flyinghighjets

    You do know that Twitter isn't just used on Mobile phones right? Clearly your upset that your "Area" wasnt clear of any wrong doings...

  • WonR

    Interesting -- but a lot of potential holes in data.

    Where is data for those hate filled rants with Redneck, conservative, gun lovers, Christians, etc. ??

  • jodi_NL

    I often tweet "faggot" and the n-word (i.e.,"nigger") from my Los Angeles home and workplace but I'm having no effect it would seem.

  • JT

    Many of you seem to have missed the point that humans literally reviewed all the tweets to ensure they were in fact 'hate' tweets, not just tweets with a certain word in them. I say the study, although imperfect, does provide some insight. There is truly a cultural difference between the east and the west. That said, we should be careful about the conclusions we draw from this data. Should there be a significant cultural difference between the east and the west, hate may be expressed in different forms in each culture.

    Forget this study, what are the known centres of racism in the country? Might those correlate with these images?

  • Jay McLeod

    I agree that a normalized metric (either number of tweets or number of tweeters) might have been more enlightening.

  • Roncameron

    This  study  is  absolutely  laughable !!   Total  nonsense.    BTW,  blacks  use  the "N"  word more than  probably  any  other  group.

  • Julia Dunne

    The truth is there are mean, snarky, hate-filled people in every town, every county, state, and region in the USA. The truth is there are mean, snarky, hate-filled people in every walk of life: political, religious, corporate, blue-collar worker, stay-at-home mom. The truth is there are hypocrites who behave one way in public, whether in real life or online, and then behave in mean, snarky, hate-filled ways behind closed doors. We don't need maps to tell us this. We just have to walk out our front doors, or log in to our social media accounts. 

  • Laaces13

     I think this is a poorly done project.  It is all based on the number of hateful tweets out of all tweets in a specific area.

    From the Mashable article of the same map,

    "Hateful tweets were aggregated to the county level and then
    normalized by the total number of tweets in each county. This then shows
    a comparison of places with disproportionately high amounts of a
    particular hate word relative to all tweeting activity," the group said.

    Orange County, Calif., for example, has the highest absolute number
    of tweets containing slurs, but due to its high level of overall Twitter
    activity, the hateful tweets are less prominent, and therefore are not
    that conspicuous on the map.

    So people in the California still hate people, they just tweet more about other crap.

  • jak

    How would you scale the data? Would you not use an aggregated measurement? 

    No study is perfect but it does offer insight. 

  • No

    are people on the west half of the country not on twitter, or do they use different terminology. This is an obviously extremely biased data collection.

  • Foiled23

    Except "I hate McDonalds." or "I hate when I run out of milk." isn't really the type of hate they were going for.