Tag Archives: University of Madeira

Data, data everywhere…Monitoring Social Media During a Crisis

Post by: Kim Stephens

When it comes to social media, monitoring these networks is the biggest concern of emergency management organizations, and for good reason: the stream of user-generated content becomes a torrent the second a crisis happens. I’ve blogged about this numerous times, but I’d like to mention a couple of resources and articles that have recently been brought to my attention.

1. Social Media Monitoring Tools Wiki  This wiki has an extensive list of over 100 free and for fee monitoring tools and services. Each item is hyperlinked, and you can sort each column, even by category. Although there isn’t an opportunity for users to rate the services or provide feedback, it is one of the few places I’ve seen such an exhaustive list.  (As an aside, I’ve toyed with the idea of a user-rated website for all tools used by emergency managers. The consumer would be asked to rate their experience with the tool and discuss what they liked and didn’t like about it, similar to ratings for merchandise on Amazon.com or restaurants on Yelp.)

Of note,  most of the tools on the wiki are designed for business clients or individuals and are probably sufficient for performing sentiment analysis. However,  these same tools have distinct limitations when they are used for the purposes of gaining situational awareness after a disaster. This brings me to#2 on my list.

2.  Mark Cameron, et. al, in the paper Emergency Situation Awareness from Twitter for Crisis Management break down social media monitoring needs of emergency managers and first responders into 5 distinct areas.

The need to:

  • Detect unexpected or unusual incidents, possibly ahead of official communications;
  •  Condense and summarise messages about an incident maintaining awareness of aggregated content without having to read individual messages;
  • Classify and review high-value messages during an incident (e.g. messages describing infrastructure damage or cries for help); and understand the impact of an incident on people and infrastructure;
  • Identify, track, and manage issues within an incident as they arise, develop, and conclude; pro-actively identify and manage issues that may last for hours, days or weeks;
  • Perform forensic analysis of incidents by analysing social media content from before, during, and after an incident.

Their solution is the “Emergency Situation Awareness–Automated Web Text Mining” (ESA-AWTM) system. The paper and this power point describes the system’s functionality as well as the trial deployment with the Media and Crisis Communication team within the Strategic Communication Branch (SCB) of the Australian Government.  The ability to classify high value messages alone would be a significant improvement. Overall the system looks extremely promising, but unfortunately they do not indicate when it will be available for broader distribution.

3. Computer aided analytics is one solution, another is crowdsourcing. This paper, “Towards Real-time Emergency Response, Using Crowd Supported Analysis of Social Media”  was written by researchers at the University of Madeira in collaboration with the IBM T.J Watson Research Center. They propose an architecture for  how crowdsourcing can be incorporated as part of an emergency response system in order to “analyze and structure social media content posted by micro-bloggers and service users, including emergency response coordinators and victims.” The key to crowdsourcing, they state, is ensuring that the crowd is given appropriate tasks to accomplish, and ultimately, that their analysis is fed back into the response community’s knowledge base in a structured way. This diagram illustrates their concept:

This is just a short list, today, but there is a lot going on in this realm. If you know of other interesting studies please post them in the comments section and I’ll add them to the resource tab of the blog.