The Twitter world of #ZumaMustFall

2015 was a rough year for South Africans, punctuated by regular social upheavals that seem to have taken on a consistent rhythm to become the new normal. The last of these upheavals for the year was President Jacob Zuma’s unexpected firing of then Finance Minister, Nhlanhla Nene. The move caused the country’s economy to take a nosedive, which, in turn, lead to calls for President Zuma to step down. This call was consolidated under the banner of the #ZumaMustFall movement, which was, itself, derived from the #FeesMustFall movement that had brought our country to a standstill just a few months before.

The frustration and anger that so quickly galvanised around the #ZumaMustFall movement could only have been realised, given the events that had gone before, and which had pushed so many South Africans to their wits’ end, and shaking so many of us out of our political and social apathy. It seemed as if the country had collectively, finally, had enough (as if #FeesMustFall had not demonstrated this already) after weathering blow upon blow to the stability of our economy, international perceptions, basic service delivery, and our own conceptions of what it means to be part of the “rainbow nation”.

Zuma’s firing of Nene seemed to be the final straw. Many believed that there was little, if any, moral justification for Zuma’s actions given his role as the supreme public servant and his mandate to safeguard our country’s wellbeing. Indeed, specific statements of outrage over Nene’s firing came from a broad-base of South Africans, from church groups, to Bantu Holomisa of the United Democratic Movement, many within the Democratic Alliance, Julius Malema and Sentletse Diakanyo of the Economic Freedom Fighters, Pieter Mulder of the Freedom Front Plus, trade unionist Zwelinzimi Vavi, activist Zackie Achmat, international media such as the Business Insider and The Economist, and folks within the ANC, notably Shaka Sisulu. Even newly reinstated Finance Minister, Pravin Gordhan, offered an unequivocal rebuke of the decision. It was an act that many despised regardless of their personal agendas. However, as the scandal unfolded, party politics and racial prejudices were worked into the gaps to cleave apart a country initially united in its condemnation.

Let’s take a look at what happened in the public Twittersphere to get a feel for how things unfolded (and continue to unfold). I collected 91,870 tweets that mentioned the #ZumaMustFall hashtag or variations thereof via Twitter. This might not represent every single tweet that used the hashtag and related terms, but it represents a substantial portion.

As a reminder, we saw the #ZumaMustFall hashtag first rear its head during the #FeesMustFall crisis. During that time, calls transmuted from the Wits-specific #HabibMustFall to #BladeMustFall and finally into the blanket #ANCMustFall. With the shift to calls for the ruling party to fall, we also saw the introduction of calls for President Zuma specifically to fall (see yellow data in Figure 1).

Figure 1: Time series showing the hourly tweets volumes for the hashtags #BladeMustFall, #ANCMustFall and #ZumaMustFall during the main period of the #FeesMustFall event.

Thus, we see that the #ZumaMustFall hashtag has been around for a few months already, but it really came into its own with the firing of Nene, which many thought reflected directly on the poor judgement of Zuma. To get a feel for the scale of the movement, I compared the volumes of tweets generated around the movement, to a few other recent events (see Figure 2).

Figure 2: Daily tweet volumes for various social events compared against one another show the scale of the public response on Twitter to each event. This chart does not necessarily include every single tweet relating to each event. In some cases, we are limited in the number of tweets that we can collect via Twitter’s publicly accessible systems. Thus, these figures are rough indications of volumes only.

What we can see is that conversation around #ZumaMustFall (in yellow) was substantial. It eclipsed events such as the recent EFF march, in red,(which was mostly driven by EFF members rather than a broad-based audience, and the release of the Nkandla report in May (blue chart area). Only the conversation around the State of the Nation address earlier in the year (not shown here) and the #FeesMustFall movement were bigger. This suggests that the movement struck a nerve. It also shows that South Africans are becoming more vocal, and more willing to take to public spaces in greater numbers to voice their discontent.

The #ZumaMustFall picture is not as clear-cut. Whereas those talking about events such as SONA and #FeesMustFall on Twitter tended to be broadly aligned in their support and interpretation of the events, the #ZumaMustFall conversation turned out to be far more divisive as various constituencies waded in and sought to fracture the movement through the use of South Africans’ favourite social crowbars: race and ideology.

By connecting together everyone that talked to each other within the context of the hashtag and its related terms, we get a conversation map consisting of 33,944 unique users. We can then run a community detection algorithm on the resulting network to find out who the main groups of people were talking about the issue and what their agendas were. Figure 3 shows what the conversation map around #ZumaMustFall looked like:

Figure 3: This conversation map shows every Twitter user that mentioned the #ZumaMustFall hashtag or a related term. Each node (circle) represents a single user and the size of the node represents how influential that user is based on their ability to cause action in others in the form of retweets and @mentions. The colour of each user’s node tells us which community that user falls into based on who they interacted with.

Figure 4 shows a breakdown of the ten communities with the most users. They collectively represent 50% of all users talking about the #ZumaMustFall movement in our data, but generated 65% of all the tweets on the topic.

Figure 4: This chart breaks down the ten largest communities talking about #ZumaMustFall based on the number of Twitter users in each community. It shows the proportion of users from our entire dataset that fall within that community as well as the proportion of all tweets about that the topic that was generated by that community.

From Figure 4, we can see that most of the main communities supported the #ZumaMustFall movement. While I have not shared individuals’ info in the article for privacy reasons, after having looked through the communities to see who the main influencers were in each, and what kind of content resonated the most within each, it is clear that support for the movement came from a wide variety of South Africans.

There were two prominent “naysayer” communities though, that promoted a race-based narrative around the movement. Indeed, the single largest community (representing 9% of all users in our dataset) was actually made up of #ZumaMustFall naysayers – that is, they criticised the movement and its motives. The fact that the proportion of tweets generated by this community (7%) was less than the proportion of people in it points to a lack of robust discussion within the community itself though. Instead, we see a pre-dominance of simple retweeting rather than two-way discussion. By looking at the content that most resonated with this community, we can see that they clearly interpreted the movement in divisive racial terms, with the most popular tweets being:

The photo of a co-operative, rather than adversarial, police officer taking a picture of three marchers became the poster image for #ZumaMustFall naysayers. This photo was used repeatedly to highlight the difference in power dynamics in the interactions with police between whites and blacks. It was used to cast the #ZumaMustFall marches as whites-only affairs, even though photos such as the following ones from the #FeesMustFall protest tell a far more complex story about white protest than the racialising and polarising one above: (Photos by The Daily Vox)

Regardless, the naysayer communities readily promoted the narrative that white protestors were simply using the #ZumaMustFall marches as an excuse for a fun day out, thus dismissing the protestors’ earnest attempts to voice their frustration and anger over the state of our country as less than valid.

The theme of the #ZumaMustFall marches being the preserve of white racists was put forward again within the other major naysayer community. This community formed around ousted former EFF member, Andile Mngxitama, who painted the entire movement as racist due to some protesters reported brandishing of the old South African flag:

More subtle versions of the white racist narrative were woven together with socialist and communist ideology to dismiss the movement as the preserve of “white monopoly capital”. Julius Malema initially supported the idea of unseating Zuma:

But Malema subsequently distanced himself from the marches, stating that they were instigated by “white monopoly capital”:

Malema’s about face was probably related to the fact that Zuma’s battering of the economy was the prime instigation for the birth of the movement in the first place, and the issue is closely aligned with the interests of banks and mining companies. (As an aside, NUMSA’s Karl Cloete expands on this particular view in an editorial piece right here on the Daily Maverick.)

The evidence shows that a notable, mostly black, minority of Twitter users interpreted #ZumaMustFall as a white racist movement. However, the majority of users that spoke about the movement seemed to support it, and these users came from all races.

The second-largest community in the data self-organised specifically to promote and co-ordinate the marches. This community represented 8% of all users in our data, but generated almost a fifth (19%) of all tweets. Interestingly, the tweet volumes were not primarily generated by a few key influencers that were retweeted many times. Instead, the community was highly decentralised with tweets coming from many individuals, which is a sign of a very involved, passionate group. It is this community that likely spearheaded the momentum behind the #ZumaMustFall movement.

Communities that have been similarly vocal during past events have all been made up of passionate activists at the core of each event. For example, we’ve seen this level of “vocalness” in previous datasets from black activists centred around the Rhodes Must Fall movement, and from EFF apologists. It is quite novel to see this kind of dynamic from an entirely new constituency – that are mixed-race, broad-based activists and critics.

The other major communities were also made up of supporters for the movement. Unsurprisingly, the Democratic Alliance had a prominent presence, driving this agenda. The EFF’s relative absence after its initial agreement with the sentiment behind #ZumaMustFall is more surprising, however, this is explained in terms of the ideological distinctions already discussed above. It is tantalising to consider the path that the movement could have taken in its initial incarnation if the EFF and other socialist/communist leaning groups had more fully invested in it. So, what do we take out of this dataset? I take out a few things:

• First, the data shows that the scale of the reaction and discussion around this event was very large. Only the reactions to #FeesMustFall and the State of the Nation address (#SONA) earlier this year were larger.


• Second, the movement was broad-based with supporters from all walks of life in South African.


• Third, this support stood out despite attempts by some to frame the #ZumaMustFall marches as a white racist movement. Naysayers used two narrative elements to undermine the movement:
  
  
  
  • Race and power dynamics were used to imply that the #ZumaMustFall movement is a naïve, “picnic” movement for white kids who want to play politics;
  
  
  • Socialist and communist ideologies were invoked to shift the discussion away from genuine outrage over the actions and morals of Zuma to the more amorphous debate over the ills of our capitalist society.
  
  
  
  

Finally, from my analysis of the data, #ZumaMustFall appears to be a real, substantial movement with broad-based support and appeal. It represented the last big social shock for our country of 2015, which may continue into 2016. Indeed, the scale of these events seems to be growing, pulling in more and more diverse sectors of our country. Where we go from here will likely be decided at the next SONA in early 2016, if the pot does not boil over beforehand. DM

Photo: Protesters carry placards as they take part in a demonstration against President Jacob Zuma in Cape town, South Africa, December 16, 2015. REUTERS/Mark Wessels.