Visualizing Crisis Related Crosstalk

Two days ago I posted an analysis on the flow of news and crowd-sourced reporting during a crisis situation. It touched upon recent events here in Kampala as well as the Mumbai Attacks, and events in Iran and China. The observation was that many of the current tools that people use during these situations, aren’t designed to be as useful as they could be. But that’s okay because there’s room for dedicated emergency reporting utilities like Vine and Ushahidi that attempt to address this.

Still, the platforms that exist today have their shortcomings, there’s a lot of room for improvement which is what my post “Asynchronous Info, Disjointed Data and Crisis Reporting” was all about. A friend of mine (another Ugandan software developer) responded with a concept for an application that he sees as a way to contextualize the cross-talk during a crisis. In the picture above you can see that he uses a scatter-plot grid to aggregate information related to an event from Facebook, Twitter, Blogs and other places from around the web.

Here’s an excerpt from his comment:

While reading the article, I couldn’t help but remember the Google Wave video that demonstrated the power of collaboration across a time-stream. With collaborators (reporters) chipping in and out of the data-stream. I believe a Wave -like variant of a newstream service could be the answer, to an extent.

Of course it would need to be highly visual, kinda like wading through a datastream. The data could be grouped in verification layers or zones from “rumour” to “verified”, especially for those that call for journalistic integrity. Massive video/audio/text aggregation for “related” material could be created for reference and background information.

He then went on to create a mockup for the app he was describing. Here’s the description of the application in his own words:

The core of the system is a visual timeline in which events/incidents/background information/videos (let’s call all of these “reports”) appear linearly along the timeline at the specific moment they are reported. Each report is then distributed vertically across “reliability” layers with sections like “false”, “trusted”, or “verified”. For example, it’s very hard to fabricate a genuine riot video during a crisis, so it makes sense that all videos are at the top.

Each report is again analysed for cross-references (tweets, re-tweets, incoming links, facebook mentions, etc). The more cross-references a report has, the larger the icon size. (So you can have a very prominent false rumour and a poorly linked trusted source)

Also, each of those cross-references is then connected to the source by a line, so for example, a retweet can be traced back to the original tweet. This would avoid issues of belated retweets across different timezones. Finally, when clicked, each report icon expands to show either the full report (for tweets) or a snippet of the report (for websites)

A quick glance at this system allows the visitor to immediately filter by: accuracy, prominence, time, connection, background information and related material.

You may remember Solomon King as one of the people reporting about the situation unfolded, despite the fact that his personal safety was at risk. He’s also the founder of one of Uganda’s more innovative web companies, NodeSix. You can find out more about NodeSix in this interview I conducted with him last year and his appearance on Appfricast 3 after Facebook engineer Charlie Cheever’s workshop here in Kampala.