Riot mapping and social media
The recent riots have shown that verifying sources can be a tricky...
What do you think? Click on the 'Comment' button above to share your views on this.
One of the claims often made around real-time social media such as Twitter is that "we're all journalists now", capable of reporting the news as we witness it and being able to braodcast our reports on an equal footing with the traditional media. The truth is often very different, of course. Looking at Twitter messages tagged with #londonriots or #ukriots and it's very hard to distinguish authentic original reports from a trustworthy (or verified) source, from hearsay, second hand reports, and blind retweets. Does the photo of a "riot in Birmingham" actually come from Birmingham, for example, or is it really a photograph from a street in London taken weeks before, as described in this review of social media coverage of the riots: How a musician and a Sikh TV channel dominated coverage of the Birmingham riots.
Verifying sources can be a tricky job, so how can we make effective use of social media as a reliable news channel? One way is to follow journalists directly who are on the scene and using social media to provide live reports: Guardian journalist Paul Lewis (@paullewis on Twitter) reported the London riots using this approach throughout Monday evening for example. Another approach is to find journalists who have started curating updates from a trusted network and who also pass on news from verified sources. During the Arab uprisings, Andy Carvin of the US based National Public Radio was widely praised for pushing the boundaries of this form of journalism - see this Guardian article for more on this. Back in the UK, James Cridland (@jamescridland) has been using online maps as a way of monitoring verified outbreaks of violence associated with the recent troubles, see his blog and below.
An important thing to notice about this map is that it only records verified events. Where the news is not so serious, it may be appropriate to use crowdsourced data from large numbers of unverified sources. This technique has been used over the last couple of years on Ben Marsh's UK Snow Map, which captures Twitter updates tagged with #uksnowamp and the first part of a UK postcode to create a live map showing where it's snowing, at least, according to the Twitterati!
One way of quickly getting an idea about whether a status update has been posted by someone who claims to have witnessed an event at a specific location is to try to find out whether the update has been tagged with geolocation data. Many phones (and cameras...) now include GPS services that capture the latitude and longitude of the device, and can add this information directly to a status update (or photo) as metadata. Through so-called APIs (Application Programmable Interfaces), it's possible to search social media websites in near real-time for updates generated in a particular area (as identifed by geolocation metadata) and plot it on a map. It's even possible to build these sorts of "mashup" yourself, as for example described in Discovering Co-location Communities – Twitter Maps of Tweets Near Wherever
Another technique for checking the reliability of a social media source is to consider the network they are part of. For example, this social network map of UK journalists shows how they follow each other on Twitter:
The credentials of anyone claiming to be a UK based journalist who isn't followed by at least some of these people might therefore be in doubt. In a more general case, if a Twitter user only follows and is followed by bots and spam accounts, you might suspect their legitimacy. By reading their tweets and seeing who they communicate with explicitly, you can also get a feel for how plausible they might be as a source. And by combing the two (mapping both geolocation and social relationships) you can also start to see whether or not groups of individuals who are known to each other are tweeting about a specific event from the same location.
But that's another story...