Emergency situation awareness: Twitter case studies

The Emergency Situation Awareness (ESA) system provides all-hazard situation awareness information for emergency managers using content gathered from the public Twitter API. It collects, filters and analyses Tweets from specific regions of interest in near-real-time, enabling effective alerting for unexpected incidents and monitoring of emergency events with results accessible via an interactive website. ESA was developed in close collaboration with users to ensure fitnessfor- purpose for the tasks performed by emergency services agencies. ESA processes large volumes of Twitter data and identifies discussion threads, trends and hot topics using language models. A burst detector generates alerts for unusually high frequency words that are filtered using text mining techniques and machine learning algorithms to identify Tweets of interest to emergency managers. An overview of the ESA platform is presented along with example case studies of its use to detect earthquakes, identify bushfire events and provide all-hazard monitoring in a crisis coordination centre. © Springer International Publishing Switzerland 2014.