Authors explain data sets related to the Mike Brown shooting part 1

Jeff: When we first looked at this data set, it was exciting to see how these hashtags: #IfTheyGunMeDown, #MikeBrown, others trended right away. #IfTheyGunnedMeDown was particularly interesting because people were applying this event to their own personal lives, but when we stretched out the scope of our data collection, we looked at the longer-term development of hashtags related to the Mike Brown shooting.

What was really interesting was that people were then associating Mike Brown with Eric Garner. People on Twitter were demonstrating that they saw the connection, that there was a pattern of unarmed black men being killed by police officers. The cry for social justice was carrying through these events on Twitter.

James: Yeah, absolutely. And I think from my perspective it was interesting […] the big hashtags that have gotten attention from previous research studies and a lot of journalistic pieces that we looked over together, hashtags like #BlackLivesMatter, #AllLivesMatter, and even #ICantBreathe didn’t really exist at first. They were not present in our dashboard. We saw #IfTheyGunMeDown that had an initial spike of activity, but what really dominated in the immediate aftermath were these factual, or “boring” hashtags that describe places, proper names, and place names like #Ferguson, #MikeBrown, #MichaelBrown. Things like #FergusonPublicLibrary where people gathered.

It was interesting from a data perspective often times you clean out proper names and place names like that. Frequently we need to suppress those signals because they obscure or dominate over underlying hidden trends in the data that you want to tease out. Quite surprisingly that noise, the factual “boring” signals were actually the finding in this case. So, yes, that was a surprising start to our study and very promising, I remember in the initial stages. It was surprising in the initial discourse of Michael Brown, along with #IfTheyGunnedMeDown. I thought that was fascinating in these first stages how the noise became the signal for us when we started working through this data.

Jeff: At the very first, what we were seeing in the data was that the Mike Brown shooting was really event-oriented. Tweets and hashtags were oriented around basic facts related to the event. Just basic stuff that is Journalism 101. Persons, places, names, stuff like that. However, as time went on, the corpus of tweets became more interpretive and more thematic about social justice.