During the last couple of years, there has been a lot of talk about social media as driving forces for revolutions. Smartphones, Twitter and Facebook have been said to enable grassroots interest groups, social movements, NGOs and oppressed populations to voice their concerns, talk back to the powers that be, and to actually make a difference.
A number of uprisings across the globe during the last few years have popularly been labeled Twitter Revolutions: the civil unrest following the 2009 elections in Moldova, the Iranian election protests in 2009 and 2010, the 2010-2011 Tunisian protests against the Ben Ali regime, and the 2011 Egyptian protests against President Mubarak. During the second part of 2011, the events during the so called Arab Spring have often been bundled together with the emergence of the Occupy Movement in discussions of how “it’s kicking off everywhere”.
In the wake of these developments, a certain friction has developed in public debate as well as in new media research. Some commentators focus mainly on the positive potential for new media audiences to come together, subvert power structures, and take control over the flows of communication in society, while others are more pessimistic and underline the risk of over-emphasizing the democratizing potentials of social media and forgetting that power structures prevail and may even be strengthened by these same media.
In a study aiming to evaluate the actual power of social media for activist politics, I analyzed a dataset consisting of Tweets posted during the first 24 hours of the Libyan uprising in 2011. I looked only at tweets using the hashtag of #feb17, and only at tweets that had an addressee in the form of another twitter account. The graph below shows the communication network.
Reading this mess might not be completely intuitive (click the image for a slightly larger version), but the circles represent Twitter accounts, and their sizes reflect how much communcation flows through them. Anyhow (you can read the full paper here), I was able to identify two key groups including, first, the Twitter accounts of traditional media companies and their journalists (grey circles), and second, a small number of activist organizations and especially high-profile activists (black circles).
What we do not see, however are any obvious traces of the allegedly earthshaking grassroots mobilization going on. Neither are any government actors from any country present in the network core. If we rely on this graph, we will end up with a description of the social media discourse during the Libyan uprising as ruled mainly by news corporations and a small number of activists and activist organizations, most of these were also coming from locations far away from where the revolution was taking place.
But what happens if we put the direction of communication under closer analysis? Does the result change if we look at whether the accounts are mainly receiving, mainly transmitting, or having a balance between these two?
Well, yes it does! In the figure below, accounts affiliated with political NGOs were categorized as belonging to Activist groups (AG), and accounts where profiles clearly stated that their owners are journalists were consequently categorized as “Journalists” (J). Official accounts of news corporations were sorted as such (NC) and governmental accounts — these were exclusively European or North American — were also collected in their own category (G). The categorization is not exact. In fact, a relatively large portion of the accounts had to be coded as “Unidentified/anonymous”. But still, the results were quite interesting.
Taking the direction of communication into account, I found that:
1. Governments and news organizations were tweeted to and about. They were mentioned as topics for discussion, and received criticisms or calls for attention or help. To a much lesser extent, if ever, were they taking an active part in the exchanges that contribute to keeping the center core of the network together.
2. Activists and activist groups, on the other hand, were feeding information into the network by mentioning and replying to other participants. These actors appeared to be engaged in relationships based on giving and/or mutuality.
The main conclusion was that a mapping based simply on how much network flow passes through a given account, regardless of direction, leaves the impression that the discourse is dominated by media corporations and a small number of NGOs and particularly prominent individual activists — largely from the world outside of Libya. If, however, the direction of communication is taken into account, a different landscape unravels, where the share of North African accounts and tweets in Arabic seems to be significantly larger. Furthermore, accounts that had peripheral positions in the first graph then appear as key agents, and accounts that were at the center of that same graph are revealed to be quite passive.
The operation is elementary, next to trivial, from the perspective of network analysis. But the example illustrates the power of visualizations: The first way of visualizing the network would suggest that grassroots mobilization is a minority phenomenon in in the analyzed dataset, while the second approach supports a view of individual activists and activist organizations as the prime movers of the studied Twitter discourse.