April 20, 2024

Idavox

The Media Outlet of One People's Project

So Why Exactly is a Hashtag Popular Among Racists Showing Up Again Online

Hint: It’s not okay to be a bot.

Cody Webb

In the run-up to the 2016 election, right-wing trolls utilized botnets and automated twitter accounts to spread propaganda. This was used to devastating effect with the propagation of various conspiracy theories (pizzagate, Hillary Clinton is a cannibal, et cetera). 2020 begins with the same tactic. Over the New Years’ weekend, the explicitly racist hashtag “itsoktobewhite” exploded on Twitter. This appeared to be a dry-run for the same right-wing social media network from 2016– shortly afterwards, this same network distributed an edited video of Joe Biden making what appeared to be racist remarks endorsing white nationalism. Our research shows that the accounts that distributed the “itsoktobewhite” hashtag share a nearly reciprocal commonality with the accounts that distributed the Joe Biden video.

The recent spate of racist hashtags intent on manipulating the system has been coming from the same network of twitter users. We used data gathered about the hashtag #ItsOkayToBeWhite to construct a graph showing the propagation of this hashtag.

An algorithm made famous by Google called PageRank provides a ranking of websites in relation to other websites, in effect displaying central nodes of a network. This algorithm was applied to the network of accounts repeating the hashtag to determine this network’s central nodes and to remove irrelevant nodes. I also discovered the community structure of the network, according to the infomap method. https://www.pnas.org/content/105/4/1118

Once PageRank and community infomap have run, we are left with the most important nodes in the network and the irrelevant nodes are blackened out. This helps to confirm who the most central nodes are in the network, and in this case it was who had the most central tweets in the network , Jackposobiec, PrisonPlanet, Carpe Donktum and KTHOPKINS are the most centrally located nodes in the network, these all share a very strong reciprocity with one another, with many of the same users in the network finding their tweets palatable . This is pretty usual for the community as they are mostly in the same social circles in this information space.

Data was then gathered on the viral, out-of-context video purportedly showing Joe Biden speaking about white nationalist values. Data for this video was then overlayed on top of the #ItsOkayToBeWhite data; PageRank and community infomap were applied to both. The originator of the viral disinformation video of Joe Biden (shown in the top left, green) shares many of the same users of the retweet network of other far right celebrities present in the information space. The green network of the viral disinformation campaign is largely interconnected with the rest of the network. This much celebrity alone puts this user on par with Jack Posobiec, with a 1921 degree nodes to Jack’s 1938.

This Network is a Tweet Source Network graph of the combined datasets. We see the main three tweet source clients: green for Twitter for android, white for Twitter web app, and blue is Twitter for iPhone. Every other color is custom code, leveraging twitter api to tweet or retweet a subject.

https://labs.graphistry.com/graph/graph.html?dataset=PyGraphistry%2FE02ONXXVH4&type=vgraph&viztoken=f576ddba925b8c232beb745e378913bb75979cfd&usertag=9723133c-pygraphistry-0.9.69&info=true&workbook=498d634bd0cc2c33

Shown below is the users leveraging tweetdeck.

Based on this analysis of account behavior, our research has determined the general composition of a distribution network for right-wing hashtag and social media influencing programs. We have shown how various third-party applications like tweetdeck have assisted automating the propagation of these hashtags and memes. We have further identified prominent accounts that serve as central nodes of this distribution network.

Webb is also the author of the Troller Report.