Processing social media data for crisis management in Athena

During a crisis citizens turn to their smartphones. They report what they see, they comment on other’s reports, they offer their help, support and sympathy and, in doing so, they create vast amounts of data. Meanwhile, law enforcement agencies (LEAs) and first responders including humanitarian relief agencies are desperately trying to improve their own situational awareness, but can struggle to do so, especially in places that cannot be easily, quickly or safely reached. Since this user-generated content is often posted to social media, LEAs can tap into these resources by analysing this data. However, making sense of this data is not straightforward. In this paper we present a system that is able to process and analyse this data through categorisation and crisis taxonomies, classification techniques and sentiment analysis. This processed data can then be presented back to LEAs in informative ways to allow them to enhance their situation awareness of the current crisis.