Using DOLLY to search for all geotagged tweets in North America between June 2012 and April 2013, we discovered 41,306 tweets containing the word ‘nigger’, 95,123 referenced ‘homo’, among other terms. In order to address one of the earlier criticisms of our map of racism directed at Obama, students at Humboldt State manually read and coded the sentiment of each tweet to determine if the given word was used in a positive, negative or neutral manner. This allowed us to avoid using any algorithmic sentiment analysis or natural language processing, as many algorithms would have simply classified a tweet as ‘negative’ when the word was used in a neutral or positive way....Only those tweets used in an explicitly negative way are included in the map.

All together, the students determined over 150,000 geotagged tweets with a hateful slur to be negative. 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.

Hmm, really, LA has none whatsoever? Weird. RD: based on methodology, this could be erroneous data given the complete dearth of data from LA, LV, ABQ, or a half dozen other western states.

For example, Orange County, California has the highest absolute number of tweets mentioning many of the slurs, but because of its significant overall Twitter activity, such hateful tweets are less prominent and therefore do not appear as prominently on our map.

It could also just indicate that western staters don't geotag, don't use twitter as much, or that the "negative/neutral/positive" algorithm they're running was flawed.

This map includes ALL geotagged tweets for each of these words that were determined as negative. This is not a sample of tweets containing these words, but rather the entire population that meets our criteria. That being said, only around 1.5 % of all tweets are geotagged, as it requires opting-in to Twitter's location services. Sure enough, that subset might be biased in a multitude of ways when compared with the the entire body of tweets or even with the general population. But that does not mean that the spatial patterns we discover based on geotagged tweets should automatically be discarded - see for example some of our earlier posts on earthquakes and flooding.

Very interesting, but without a lot more insight into methodology, I'll be having a truckload of salt to go with this.

They address that a bit in part of the post that I didn't quote before:

Yeah, their FAQ acknowledges that potential source of bias:

Totally agreed. Just thought it was an interesting east/west division that makes me go, "Hm. I wonder what's going on here?"

Almost all of the problems you guys are finding with this data are addressed if you simply read the two paragraphs that go along with it. There's no algorithm to mess things up, each tweet was looked at individually to determine if the word was used in a positive, negative, or neutral manner. It's normalized by overall the overall number of tweets in each county so that population density (or the popularity of twitter varying between regions) is not a factor. If you actually look at the links, this is all explained in detail (including all the words that are used, their variations, which ones had to be left out, etc.)

This could then be disproportionately representative because the tweets were not looking in the southwest for Spanish-language hate speech, for higher populations of individuals identifying as homosexual, or maybe just because racists are more likely to be twitterers in the eastern US than the western (not that any of these are concrete, I'm just speculating). Any of these would confound the results enough to be useless.

Khross, I'm genuinely confused as to what you mean. What demographic data do they need to make their information useful?

Khross, yeah, that would make it a lot more useful for making statements about the population as a whole, as opposed to just twitter users. I still think it's interesting given that limitation.