Abstract

Few journalists can reach the remote regions of Syria. Instead, thousands of citizens are helping get the news of the devastation out.
But three years in and with no end in sight, Syria Tracker’s task to report news from the conflict has only grown. It is now believed to be one of the longest running crowdsourcing projects in the world, mapping over 4,000 geo-tagged verified eyewitness reports. It also incorporates large-scale data-mining techniques, so far scanning more than 160,000 news reports and 80 million social media updates. Maps (like the one created for Index on the next page) show the scale of citizen journalism happening in Syria, and its impact.
The Syrian conflict began in March 2011 when localised protests against President Bashar al-Assad’s regime started to spread nationwide and, following military intervention, turned into an armed rebellion. The country soon descended into civil war, and both sides have been accused of serious human rights abuses. The death toll is now thought to have surpassed 100,000, with some sources putting the figure closer to 150,000 (including an estimated 11,000 children). Reporting the war is one of the world’s most dangerous jobs. Aside being caught up in the violence, journalists also risk kidnapping. Many foreign news associations have withdrawn their reporters from the country. Citizen journalism has become a crucial way of informing the world about what’s happening on the ground.
Taha Kass-Hout, a US-based social scientist and Syria Tracker’s founder, says the work is relentless, “like a hurricane is happening every minute”. Each member of the core team works around two to three hours a day on the project, somehow also holding down day jobs (as doctors, journalists, professors, PhD students and computer scientists). For the first two years, team members struggled to keep their identities secret and received constant threats, sent to their personal emails and Facebook pages. “We were hacked; the Syrian Electronic Army constantly tried to shut us down; but finally we ‘came out’ to show the project is manned by real people, not just a computer algorithm,” says Kass-Hout.
Protecting the anonymity of their sources is, however, non-negotiable. The team immediately discounted the possibility of using text messages, with networks being monitored by the Syrian government. To power their project, they chose to use Ushahidi – open-source software for mapping and data-harvesting – which was developed after the 2008 Kenyan election and earlier this year helped with the public search for the missing Malaysian plane.
Only verified data is published – which equates to about 6 per cent of what they receive. Data-mined reports first go through an algorithm to look for patterns, before being analysed by members of the team. The crowdsourced reports are all manually checked – a process that can take several hours or several days. Any information that might identify the subjects is deleted; date stamps, landmarks and other nearby reports are correlated. Much of the video content is gruesome and of shaky quality, making the task even harder.
On the ground, the group has more than 600 citizen reporters. They aren’t employees, but around 12 have worked with Syria Tracker constantly since 2011. “We’ve had sources that have worked with us for months, then suddenly the contact drops off. We never know if their group was dismantled, they’ve been killed or they’re just missing.”
The longevity of the project has also added complications. “Crowdsourcing projects often slow down because people get fed up that you aren’t sharing anything back. So we make a point of engaging, sharing feedback, sending maps. Even if just to say, ‘Hey, your info has been used here.’” Kass-Hout pinpoints their greatest achievement as when government agencies, NGOs and the media started to take their data seriously, publishing it alongside numbers from the UN.
But the “numbers game” is also frustrating, says Kass-Hout. “People query the number of deaths [currently over 100,200 documented by Syria Tracker]. They say the data is not representative, but it is clearly representative of something. Even rumour is valid. There is no such thing as bad data.” One of the issues Syria Tracker has highlighted is the rising number of reports of women and children being attacked. “Even if you ignore the specific numbers, you can see that reports of female deaths have hit as much 18 per cent of the total deaths since the crisis began. And these are reports of rape and torture. This is not collateral damage.”
Only verified data is published – about 6 per cent of what they receive
Syria Tracker insists all it is trying to do is provide another tool for those attempting to piece together the full picture. “This is not a clinical trial,” adds Kass-Hout. “We are telling a story, it’s a living record.” And so it continues.
