Improving Social Media Monitoring and Analysis Tools for Emergency Management

Open source data from social media platforms is becoming a critical source of intelligence to inform situational awareness in the emergency management domain. However, the use of social media data in emergency management organisations is not yet widespread. This paper reports on findings from four interviews with social media subject matter experts from Australian emergency management organisations. The first goal of the interviews was to understand how social media monitoring and analysis tools are used to meet the information needs of emergency services organisations. A second goal of the interviews was to capture their insight as to how such tools could be improved, to inform the ongoing development of CSIRO’s Emergency Situation Awareness (ESA) tool. The ESA tool provides crowd sourced information in near-realtime from Twitter about all-hazard types for emergency managers. ESA collects Tweets from Australia and New Zealand and processes them to: identify unexpected incidents, monitor ongoing emergency events and provide access to an archive to explore past events (Power et al. 2014). This paper reports on the findings from these interviews. All four interviewees were using a range of social media monitoring and analysis tools such as Geofeedia and Signal as well as the ESA tool. Our analysis revealed that participants valued the following characteristics when analysing social media data for emergency management: saving time, triangulation of multiple sources of information, “More eyes on the ground”, geographic and temporal context, information about individuals, collaborative tools and cost minimisation. However, as one participant noted: “there is no one product that suits all our needs”. We spoke to participants about how the current social media monitoring and analysis tools could be improved. The key features they sought were: 1. Automated evaluation of information sources 2. The ability to follow social media content from specific individuals 3. Geographic representation of search results 4. Access to real-time feeds across multiple platforms, and 5. Flexibility to retrospectively modify searches. We conclude by describing how these recommendations have been used to improve the delivery of CSIRO’s ESA tool.