Our first case explores the relationship between social media and social justice movements, as social media and mobile streaming applications provide a potent form of storytelling power to users across the communicative landscape. How do social media platforms, such as Twitter, affect public discourse about social justice? Notably, perhaps, social media platforms have the potential to change the relationship between news media and the public in significant ways, as virtually everyone now has the ability to document and livestream events to a global audience. As noted earlier, social media has become a primary venue for public commentary about current events and has disrupted some of the gatekeeping power once held by national news outlets and talk radio in the discussion of public affairs.

Some of the most poignant examples of this restructuring of communicative power can be seen in social justice movements and the instant release of imagery and commentary in the wake of multiple shootings of Black men by police officers across the U.S. in recent years. For instance, Diamond Reynolds livestreamed the moments after the shooting of her fiancé, Philando Castile, when they were pulled over by police for a broken taillight in Falcon Heights, Minnesota.1 Videos were posted online when police in Baton Rouge shot Alton Sterling, prompting an investigation from the U.S. Justice Department. As mentioned earlier, civil unrest followed the shooting of Michael Brown, an unarmed Black teenager in the St. Louis suburb of Ferguson in the summer of 2014. As the hashtag #Ferguson trended on Twitter, national and international news outlets followed social media activity in covering the protests, looting, and militarized police response. And in Cincinnati during the summer of 2015, Sam DuBose, an unarmed Black motorist, was shot and killed during a traffic stop by a University of Cincinnati police officer. Afterwards, local community groups led by @BlackLivesCincy and @theIRATE8 quickly mobilized on social media to decry the incident and confront competing narratives that it was justified.

In this chapter we engage in a big-data analysis of #BlackLivesMatter and related hashtags that were invoked on Twitter in the aftermath of the Mike Brown shooting in the St. Louis suburb of Ferguson in 2014. The purpose of this chapter is to present not only a broad overview of Twitter activity and viral hashtags after the event, but moreover, to build off our previous examination of affective discourse within the millions of tweets that defined the social movement related to the killing of unarmed Black men at the hands of police officers.2 The findings presented here show how social justice groups in general, and the public in particular, use social media to provide a more diverse array of commentary about the meaning and implications of civic activity, and indicate that historically marginalized groups and the broader public have exercised their First Amendment rights in ways that have redefined the relationship between public communication, national news outlets, and international networks. Additionally, our analysis of tweets in the aftermath of the Mike Brown shooting and the non-indictment of Darren Wilson show that individual posters tended to relate the meaning of these events to their own lives and framed these events as relatable to a broader array of personal experiences and events.

Deen Freelon, Charlton McIlwain, and Meredith Clark suggested that the internet in general, especially Twitter, were instrumental in developing the Black Lives Matter movement. Although the #BlackLivesMatter hashtag was generated in the summer of 2013 after George Zimmerman was acquitted for the murder of Trayvon Martin, according to Freelon, McIlwain, and Clark it was not a popular one until August 2014, when it was frequently invoked during the Ferguson protests.3 Other studies have further detailed how social justice movements have effectively engaged social media in general, and Twitter in particular, to provide counternarratives to legacy news media, to intensify public debate and criticism about law enforcement, as well as to yield more in-depth dialogue on and personalization of the issues.

Although their study was not specifically related to the Black Lives Matter movement, Sarah Jackson and Brooke Welles showed that minority voices have used Twitter to establish effective counter-narratives challenging police activity that extend into mainstream media.4 Ryan Gallagher, Andrew Reagan, Christopher Danforth, and Peter Dodds found in their analysis of 800,000 tweets the #BlackLivesMatter hashtag amplified public criticism of police killings of unarmed Black men.5 Moreover, users invoking the #BlackLivesMatter hashtag tended to have more “informationally rich conversations” than their counterparts who used the #AllLivesMatter hashtag.6 Those posters using the #BlackLivesMatter hashtag demonstrated more diversity in word usage and cut across topic networks in comparison to tweets featuring the #AllLivesMatter hashtag. Alexandra Olteanu, Ingmar Weber, and Daniel Gatica-Perez reached a similar conclusion, showing that Black people who use the #BlackLivesMatter hashtag are more likely to engage the ethical dimensions and personal implications on sensitive topics.7

Freelon, McIlwain, and Clark analyzed tweets from a year-long time period (June 1, 2014 to May 31, 2015), which included events over two months before the Michael Brown shooting, and nearly a year after, and other studies have explored the #BlackLivesMatter phenomenon on social media more broadly.8 However, the examination presented here focused on tweets from specific time periods related to the Mike Brown shooting (the immediate aftermath of the shooting itself, and the non-indictment of Officer Wilson) when interest and emotions were arguably highest, to better understand the story elements that individuals tend to tweet about for a defined event. The purpose of Freelon, McIlwain, and Clark’s research was to study the Black Lives Matter organization, which used the #BlackLivesMatter hashtag, but the hashtag itself was not the focus of the study. Additionally, their report covered Black Lives Matter’s involvement in an array of cases, including ones involving Eric Garner, Mike Brown, Walter Scott, and Freddie Gray.9 However, our data presented in this chapter focus on an array of hashtags for just two distinct time periods related to a single event (the immediate aftermath of the Mike Brown shooting, and after the non-indictment of Darren Wilson three months later).

The time periods noted here are ripe for analysis of the role of social media in social justice movements. Social media provided instantaneous imagery and commentary in the civil unrest that followed the shooting, and moreover, as the hashtag #Ferguson trended on Twitter, national news outlets seemed to be following social media activity in covering the protests, looting, and militarized police response. It is clear based on the cases presented in Guatemala, the Arab Spring, Occupy Wall Street, Black Lives Matter, and Ferguson, as well as feminist, trans, and more recent LGBTQ movements in China that social media provided a powerful platform for previously unheard voices. Papacharissi explained this kind of phenomenon as affective expression in the form of networked publics that “want to tell their story collaboratively and on their own terms.”10 Moreover, these “affective publics” tend to “produce disruptions . . . of dominant political narratives by presencing [sic] underrepresented viewpoints.”11 From this review of literature on social justice movements, social media have presented significant opportunities for the disturbance and redirection of dominant and oppressive narratives.

To better understand the “affective publics” described by Papacharissi for the social network analysis in this study, we adapted Robert Entman’s approach to studying media frames that understands how certain “schemas” convey meaning through the selection of and emphasis on certain story elements.12 Based on the national and international attention about the role social media (especially Twitter) played in the aftereffects of the events of Ferguson, the overarching goal of our research project was to understand which story elements people were focusing on in their tweets and hashtags, as well as how these platforms gave voice to underrepresented publics.

While previous research has broadly explained how social justice movements have effectively engaged social media in general and Twitter in particular to provide counternarratives to legacy news media and to intensify public debate and criticism of law enforcement apparatuses, our examination evaluates the most used and impactful hashtags in the immediate aftermath of a specific event (the Michael Brown shooting in Ferguson). We also examine changes over time, including triggering events where the hashtags originally invoked during the immediate aftermath of the shooting might spike again, such as the legal decision to not indict police officer Darren Wilson for the killing. Specifically, we asked the following sets of research questions:

What were the most used and impactful hashtags in the immediate aftermath of the Mike Brown shooting in Ferguson? Moreover, what were the broader social and political takeaways from this event and subsequent social media activity? What was the broader impact on the public sphere and social discourse around Black rights and the racial issues with law enforcement? This set of questions entails both data analytics and a qualitative component. First, we wanted to know which hashtags went viral and were used most frequently in tweets. Second, we wanted a qualitative assessment of the meaning and implications of those most popular hashtags. What kinds of messages (or frames) did they convey? Did they focus on places and names in the story (as an event), or did they provide commentary, a call for action, or something else?

We also wanted to examine change over time by analyzing what we imagined would be triggering events where the hashtags (invoked during the immediate aftermath of the shooting) might spike again, including legal decisions such as Darren Wilson’s non-indictment. We wanted to know which were the most used and resonate hashtags after the non-indictment, which again requires both data analytics and qualitative analysis: which hashtags were used most frequently after Darren Wilson was not indicted, and what were the meaning and implications of those hashtags? Rather than understanding social media activity as being limited just to the events of Ferguson, our analysis showed that posters were relating what happened in Ferguson to other racial issues involving unarmed Black people and police.

As described in Blevins, Lee, McCabe, and Edgerton, we extracted every tweet from across the globe posted in the four months after the Michael Brown shooting from the open-source Twitter historical archive related to #BLM, and created a network showing which users and regions responded to one another during that intense period from August 2014 (the month of the Michael Brown police shooting) to December 2014 (the aftermath of the Darren Wilson non-indictment).13

We developed an automated extraction process in Python 2.7 to extract the Twitter data, and search the data for a number of specified terms, from hashtags to words in tweets. The program stores these results by day as a flat JSON file, formatted in two ways, with one data structure for exploring tweet-retweet relationships, and another data structure for viewing basic descriptive statistics about the search terms.

The tweet-retweet relationship data structure builds arrays of nodes and links based on the Twitter historical archive. Nodes consist of users in the searched data, and links are built between those users and others who retweeted them. In this process, we preserve important data, such as tweet text and time of tweet, for more detailed exploration. These nodes and links are visualized using a modified D3.js force-directed graph that is filterable by time.

The descriptive statistics data structure creates elements for every search term in every tweet and groups each of these elements according to several parameters (i.e., retweet vs. original tweet, time of day, day of week). We built a data visualization dashboard with the crossfilter.js and dc.js libraries for filtering and visualization. The descriptive statistics and visualizations allow for interactive explorations of these data:

Note that in Data set # 1 the most popular hashtags are proper names of the victim (e.g., #MikeBrown and #MichaelBrown). Other place and proper names used in hashtags (e.g., #FergusonPolice, #DarrenWilson, and #EricGarner) were less interesting in our view, as they merely referenced a basic element of the story (a person or a place). Hashtags such as #AllLivesMatter, #BlackLivesMatter, and #JusticeforMikeBrown can be considered as ideological markers because they indicated a particular position (or belief) about the event. #JusticeforMikeBrown is the most popular of these, and it also includes the proper name of the victim in the hashtag. We considered hashtags, such as #IfTheyGunnedMeDown and #ICantBreathe as conceptual markers because they make personal conceptualizations of (or references to) the story. Interestingly, #IfTheyGunnedMeDown was the most popular of these in the early weeks after the shooting. In a short amount of time, people were making the shooting a more personalized issue, rather than referencing it as a separate, single event.

In Data set # 2 we found that hashtags with proper names (e.g., #MikeBrown) were most prominent in the first wave of Twitter activity and are the first hashtags that go viral. However, more conceptual markers, such as #IfTheyGunnedMeDown, and ideological tags, like #BlackLivesMatter, register only faint activity in the immediate aftermath of the Michael Brown shooting.

By November of 2014 (a few months after the shooting), in Data set # 3 and near the announcement of Darren Wilson’s non-indictment in the shooting, more conceptual markers, such as #IfTheyGunnedMeDown and #ICantBreathe, which include personalization of the issues, dominate the Twitter network along with #MikeBrown. We examined change over time by analyzing triggering events where the hashtags spike again (e.g., Darren Wilson’s non-indictment) and the Eric Garner death. By this point, conceptual hashtags like #BlackLivesMatter and #ICantBreathe have eclipsed #MikeBrown.

Furthermore, in Data set # 4 #IfTheyGunnedMeDown is a far more popular hashtag at times and does not always share links to #MikeBrown. Here the two markers appear to be independent with little to no correlation in their activity. As ideological hashtags such as #BlackLivesMatter and #ICantBreathe eclipsed #MikeBrown, the #EricGarner marker appeared to have a multiplier effect – amplifying activity on the Ferguson-related hashtags. Several months after the Michael Brown shooting, and in the wake of the legal decision to not indict Darren Wilson, the hashtags are much more interconnected and correlate to one another.

Textual Analysis of Hashtags

Based on the unexpected dominance of conceptual tags such as #IfTheyGunnedMeDown, we did a qualitative textual analysis of the use and significance of some of these hashtags. One of the more notable elements of these hashtags is that they included first-person personalization of the issue. Using the hashtag #IfTheyGunnedMeDown, individuals juxtaposed two dissimilar images of themselves: one, a wholesome picture of the individual, perhaps attired in cap and gown at a high school graduation; the other, the same person in street attire, maybe holding an alcoholic beverage or cigarette. The question being: if the police killed me, which picture would be in the news – the wholesome high school graduate, or the menace to society? By featuring two contrasting images of the same person, these posts demonstrated that one picture alone does not tell the whole story of a person; and questioned the tendency of news media to focus on the one image that contributes to the “menace to society” narrative.

In reaction to eyewitness accounts that Brown was surrendering with his “hands up” before being shot, several posts on Twitter using the hashtag #HandsUpDontShoot featured images of people holding their hands up. One of the most potent was a video of kids on a school bus chanting: “hands up, don’t shoot.” The message suggested that Michael Brown “could have been me,” and engages concern about police officers overestimating the threat posed by Black suspects, and too quickly responding with deadly force.

Social media appeared to change the relationship between news media and the public, as tweets and posts did more than just reiterate the images and messages from traditional news outlets about the events in Ferguson. Rather, social media was the platform for people in Ferguson to document what was happening to a global audience, and the primary venue for public commentary. For instance, the conversation from (and about) Ferguson reached as far as the Middle East, where Palestinians tweeted in solidarity about racial injustice.14 Several players for the then-St. Louis Rams attracted international attention when they came onto the field before a National Football League game imitating the #HandsUpDontShoot thread on Twitter.15 Social justice advocates were able to help drive the local, national, and international conversation through social media.

From this broader qualitative examination, it appears that social media provided a forum for both a community in Ferguson and the public at large to tell its own stories in the aftermath of the shooting and challenge the images that tend to pervade national news. In a mediated world dominated by national outlets, social media allowed the public to exercise its First Amendment rights in a way that changed the balance of communicative power and enhanced everyone’s ability to relate the meaning of the events in Ferguson to their own personal lives.

The data visualizations presented here bring shape to our understanding of social movements and political action as it plays out on social media. For instance, our network visualizations provide a more visceral form of what a “social movement” may look like as it develops on social media, compared to more conventional appearances in terms of strikes, protest marches, and sit-ins.

Moreover, the visualization of movements taking place on Twitter can reshape our understanding of how political action takes place in the digital era. We have used network analysis techniques to track how social justice hashtags attain a “viral” status, and have found that factual and descriptive hashtags, including proper names such as #MikeBrown and place names such as #Ferguson are the first wave of hashtags that become viral. More conceptual and ideological markers, like #BLM, registered only faint activity in the immediate aftermath of the Michael Brown shooting. But after one to two weeks, they dominated the discourse and captured more media attention. In examining change in the hashtag behavior over time by analyzing triggering events where the hashtags register large increases in activity, legal decisions such as Darren Wilson’s non-indictment, and the death of Eric Garner, appeared to have contributed to greater personalization of the events.

The results of this study support Papacharissi’s explanation of social media activity as an affective form of expression for groups and individuals about social justice issues.16 By telling their own stories, on their own terms (as indicated by the conceptual hashtags) these “affective publics” disrupted the power typically held by mainstream news outlets, and in the process, changed the conversation from one that focuses on basic story elements (people, places, and events) to one in which the meaning of the event is more internalized (e.g., #IfTheyGunnedMeDown and #ICantBreathe). What is unique in the aftermath of the Mike Brown shooting is that the most meaningful hashtags were the ones that helped to frame the shooting as something relatable to the posters’ own lives and experiences. These conceptual hashtags framed the event as something personal – if “I” were gunned down, “I” can’t breath, and so on. Twitter users did more than just use the platform to simply comment on an event (#MikeBrownShooting) that occurred at a particular place (#Ferguson); and even more than reinforcing popular ideological frames (e.g., #BlackLivesMatter v. #BlueLivesMatter). Rather, the conceptual frame presented in the #IfIWereGunnedDown hashtag was more dynamic – as it personalized the issue for both Black and white individuals. In the juxtaposition of images of single persons, Black people showed that police and legacy media tend to unfairly characterize them based on appearance, whereas whites demonstrated an awareness of their own privilege – that in similar situations, police and media do not make the same assumptions. The employment of personal conceptual frames in hashtags is ripe for further qualitative analysis in other cases.

For advocates of social justice, we would also caution that social media, as a platform, is not just about liberating the voices of the marginalized. Papacharissi’s explanation of affective expression of networked publics can also be applied to hate groups.17 While social media has helped social justice advocates to be more effective storytellers, it also empowers hate groups and others who use these digital tools as forms of intimidation through trolling, cyberbullying, and social media mobbing, in which targets are relentlessly barraged with insults, threats, and vulgar memes intending to drown out more respectful voices in the process.18 How social media can be used to disrupt social justice efforts, empower hateful expression, or practice intimidation is also deserving of further scholarly attention from communication researchers.

Furthermore, the set of visualization and text-mining tools on social media data employed in this study can transcend social justice applications, and we envision that our social network analysis method can have broad applications across disciplines. In developing the project presented here and its results, we make the machine learning algorithms that we have applied to the Twitter archive, as well as the visualizations we developed from the data, accessible on a cloud platform as online research tools for scholars and students to analyze social justice hashtags and the social media discourse at a big-data scale.

While the data presented in this chapter shed light on how discourse about social justice takes shape on Twitter, it does not address the context of the on-the-ground efforts of social justice advocates and how social media is but one part of their media strategy. Our next chapter explores some of the historical background of social justice movements in their struggle and relationship with legacy news outlets to deepen our understanding of social media’s significance in the present landscape.