One of the biggest challenges news organizations face is the real-time aspect of newsgathering: the massive problem that is making sense of the torrent of information that floods in when breaking-news events take place. How do you process, and then verify, and then organize, hundreds and often thousands and sometimes millions of discrete data points, even before those points transform into something that resembles useful information?
A team at Ushahidi, a 2009 Knight News Challenge winner, has been working on that problem since early 2010, developing a way to help users manage large quantities of data more efficiently — and to help verify and prioritize information as it flows in from the crowd. That system, Swift River, has just won a $250,000 Knight News Challenge grant to focus on the verification aspects of crowdsourced data. The project, its director, Jon Gosier, has said “is all about adding context to content.”
The next phase of Swift River will be to figure out ways not just to filter information provided by the crowd, but actually to make the crowd into the filter itself. The platform, a collection of APIs, already tracks user behavior to determine the accuracy of incoming content. (For more on how it works and what it looks like, check out the video above.) “Basically, we have applications that allow the user to create a baseline of what they consider to be trusted information,” Gosier told me — the River ID API, for example, which establishes user identities tied to mobile phones, Twitter, and web feeds. Swift River uses that to create approximations of trust, crowdsourcing not just information, but verification itself. “We learn from what [users] do in our system,” Gosier notes, “and then we can start to match future information against that.”
Swift River, like Ushahidi, had its genesis in crisis situations — but, like Ushahidi, it has evolved into something much bigger and, Gosier thinks, more broadly useful. The problem of real-time verification, he notes, “is a problem that affects more than just the humanitarian community. And so, to solve that problem for them, we had to think about, ‘How do we solve the problem for everyone?'”
And the answer involved thinking in terms of the atomic unit of crisis response: information itself. “Our primary use case — our primary idea of a test user — was the journalism and news community,” Gosier says. “Because, obviously, they have a huge stake in a platform that can verify and filter out misinformation.”
The Knight funding will pay for dedicated workers to build out that platform. Swift River is fundamentally a tech product, and requires technological expertise to build out. “Thus far, we’ve cobbled things together as best we can,” Gosier says, but the team needs engineers who can work on the particular problem of capturing user history in order to assess informational accuracy. “That’s the biggest, most computational, most complex, and outstanding component of the system that has yet to be built,” Gosier notes — “and the Knight funding will be applied to that.”