The protest on the Shahbag square in Dhaka is a remarkable show of an independent, spontaneous, non-partisan movement erupting out of pain silently carried by the people of Bangladesh for decades. It is an admirable non-violent, tenacious, proud and confident protest in the real world. The protest has been partially organised online and echoes across social media. This article visualises the emerged online conversation. This is best done with Twitter since it makes data more readily available to third parties than for example Facebook, and people freely and frequently express themselves in tweets.
Before we dive into Twitter let us take a look at the bigger picture. The significance of the current demonstrations on Shahbag can be visualized by a long-term view of the war crime topic. Google Trends visualizes the relative popularity of search terms over time. We can compare the search volume of shahbag, bangladesh war crimes, rajakar, and jamaat-e-islami over the last three years. There is steady searches for jamaat-e-islami and rajakar and a more recent interest in bangladesh war crimes and shahbag. (Latter is probably attributable to a restaurant with that name.) Most interestingly we can identify a sharp spike across all four terms in February 2013 coinciding with the recent war crime judgement. The ‘A‘ in the graph identifies the time of the Washington Post publishing a story about the trial and sentencing.
Twitter, of course, is a great place to eavesdrop on the crowd’s conversation and mood. The above image visualizes keywords of a thousand tweets with the term shahbag (from 10/02/2013 and time shown is GMT/London time so add 6 hours for Bangladesh local time). It illustrates a rise of conversations after noon and peaking after 1:30 pm (Dhaka time) with conversations focusing on the Bangladesh national team joining the protest. You can yourself explore the latest conversation trend around shahbag (the website uses a Java plugin and takes a bit to load so it may not work for everyone).
There are some amazing tools available to look deep into the stream of conversations. Socialcollider, inspired by data analysis for particle colliders, visualizes the stream of conversation over time and across topics. A search for shahbag indicates that the topic started trending less than a week ago. It crosses various subjects, some related and some tangent, for example, innocent, daily conversations which are suddenly drawn into the shahbag conversation with the emergence of the topic. Additionally striking is that the news feeds by AP, CNN, NYTimes and others barley touch the topic. In the top-center we can see a current, tight interaction and conversation about the topic happening with frequent cross communication, re-tweets and mentioning. You can explore it for yourself at Socialcollider.net, just select phrase and type shahbag or any other terms and press collide!. You may need a bit of patience though since the tool has to execute a few queries in the background for you.
Finally, Revisit offers a view identifying individual Twitter accounts by their avatar image scaling them by importance, i.e. retweets and mentioning. It gives a great overview of the flow of tweets, the individuals involved, and accessible, simple analysis of who has an impact. You can try it yourself (I found it not to work with the latest Chrome browser).
The exciting aspect of following a popular movement online is the social aspect. There are fast moving conversations appearing from nothingness followed by cross-pollination with other topics. Finally, what makes the streams of opinions, which express a superficial dissonance, a movement is an underlying harmony greater than the individual discords. Indeed a reflection of the real world.
Christian Prokopp is a Data Scientist in London, UK, and focuses on big data, cloud computing, and machine learning. Christian has a BSc from Germany, a MCom and a PhD from Australia. You can follow him on Twitter @prokopp, read his blog and all his ClickIttefaq columns.