MORE than 800 people have been killed in Egypt since police attacked protesters supporting former president Mohamed Morsi, who was deposed on 3 July. Could analysing Twitter have helped to avoid the massacre?
In early 2011, when Egyptians protested in Cairo鈥檚 Tahrir Square and overturned Hosni Mubarak鈥檚 regime, social networks like Facebook and Twitter got a lot of the credit for helping rebels spread their message.
Now, just over two years later, Egypt has been plunged into violence. Twitter isn鈥檛 in the spotlight in the same way as it was during the Arab Spring, but nonetheless it did show that the latest conflict was coming.
Advertisement
Ingmar Weber and colleagues at the have created the Political Polarization Index to measure tension on Twitter. They looked at Egyptian tweets between March 2012 and June 2013 and assigned each user a score between 0 and 1 depending on which prominent figures 鈥� Islamist or secular 鈥� they had retweeted. The 鈥減olarity鈥� of popular hashtags used to group messages was then calculated by averaging their use across all Egyptian tweeters, taking each user鈥檚 religious preferences into account.
By tracking how the polarity of these hashtags changed, the team was able to see the level of political divergence between the two main groups. If secularists and Islamists were both talking about politically neutral topics like the iPhone or Justin Bieber, then polarisation was assumed to be low. But when more secular users are tweeting hashtags like #tamarrod, roughly meaning 鈥渞ebel鈥�, polarity was judged to be high, for example.
They found that the increasing political polarisation they measured on Twitter preceded real-world strife. 鈥淨uite strikingly, all outbreaks of violence happened during periods where the hashtag polarity was comparatively high,鈥� the team says.
聯Increasing political polarisation measured on Twitter preceded real-world strife聰
It doesn鈥檛 mean Twitter can suddenly predict all events, but Weber鈥檚 team hopes that such measures of rising tension might give governments enough foresight to steer away from violent conflict.
鈥淚f governments realise that society is drifting apart, they might think of positive countermeasures,鈥� says Weber. In Egypt, the tension online and offline entered a 鈥渞ed zone鈥� during the in November and December 2012, he says. That could have indicated to the Morsi government that it should reconsider its actions, especially as tension didn鈥檛 really drop again, even if the streets were quiet at the time.
Weber says they might improve the system by keeping track of whether individuals have used polarising hashtags before, as a measure of whether discontent is on the rise. 鈥淚f 100 users use an anti-Morsi hashtag, it might matter whether they are just 鈥榯he regular suspects鈥� or are users who have not been politically active in the past but have now decided to express their frustration,鈥� he says.
Christopher Neu of Techchange in Washington DC, which trains people to use technology to drive social change, agrees that insight from social media can be useful in conflict situations, but says that it would function like a canary in a coal mine 鈥� warning people of impending crisis, but not necessarily helping them fix it. 鈥淏eing able to tell that something is wrong and doing something about it are very different,鈥� he says.
This article appeared in print under the headline 鈥淚 predict a riot鈥�
