Computer technology that can mine data from social media during times of natural or other disaster could provide invaluable insights for rescue workers and decision makers, according to an international team writing in the International Journal of Emergency Management.
Adam Zagorecki of the Centre for Simulation and Analytics, at Cranfield University, Shrivenham, UK and David Johnson of Missouri State University, Springfield, USA and Jozef Ristvej of the University of Zilina, Zilina, Slovakia, explain that when disaster strikes the situation can change rapidly. Whether that is during flooding, landslide, earthquake or terrorist attack, understanding the complexities of the situation can mean the difference between saving human and animal lives, reducing environmental impact and preventing major economic loss.
The team points out that advances in information technology have had a profound impact on disaster management. First, these advances make unprecedented volumes of data available to decision makers. This, however, brings with it the problem of managing and using that data. The team has surveyed the state of the art in data mining and machine learning in this context. They have found that whereas earlier applications were focused on specific problems, such as modeling the dispersion by wind of plumes — whether from a chemical plant leak, fire or nuclear incident — and monitoring rescue robots, there are much broader applications, such as constructing situational awareness and real-time threat assessment.
Data mining during a disaster can pull in information from unstructured data from news reports, incident activity reports, and announcements, as well as structured textual data from emergency services, situational reports and damage assessment forms. In addition, it can utilize remote sensing data, as well as more frequently now, mobile phone images and video, and satellite and aerial images.