People are contributing digital data at all times everywhere. Some of this data is found on social media platforms: consider the constant stream of user-generated posts on Twitter, Facebook or Weibo. Mobile phone networks also transit or host a plethora of user-generated data, from call data records to SMS text messages sent to a popular short-code. Increasingly new platforms are emerging that allow users to participate in specialized structured or formal data generation and real-time reporting. For instance, the ELMO system from The Carter Center allows a host of trained field observers to submit focused formal reports that help to monitor a country’s ongoing national election.
The sources, purposes and particularities of this user-generated data seems to be forever growing, often created around seemingly the most mundane and banal of topics. But this torrent of user-generated data is also produced at the most serious of times and in the most critical of places: during moments of national crises, political debate, violent attack, natural disaster, or civic protest. It is during these moments – during the witness of some critical event – that the deluge of user-generated data can be harnessed towards a critical and even life-saving analysis and response.
A number of systems already exist with an aim towards applying user-generated content towards positive resolutions during critical events. For instance, the crisis mapping community has developed real-time geo-locating systems that can help first responders locate and attend to people in immediate need. In general, however, these systems are media source specific – they work with a specific media type (for instance specialized SMS text messages or posts to a specialized website, or just with Twitter or Facebook). Focusing on a single media type can often result in beneficial results, but in other circumstances leaves a wealth of potentially central data untouched.
In our work we examine when and how multiple media types can come together resulting in more comprehensive data sets that countenance smarter analysis, resulting in more impactful responses during critical events. In this research we aim to invent platforms and processes to integrate and triangulate multiple media types that as a whole create actionable intelligence beyond the sum of any data parts. Twitter plus Facebook plus SMS plus ELMO plus Weibo – through integration, triangulation and thoughtful analysis – should create a richer understanding when brought together compared to just any one source on their own. We also are interested in how these approaches vary in effectiveness across different types of communities or critical events. Are there better applied to election monitoring and natural disaster response or do they work just as well during times of civic unrest or communal violence?