The team of researchers will be working with the Fairfax Fire and Rescue Department and a social media working group for emergency services at the Department of Homeland Security Science & Technology Directorate, among other groups, to refine the required information needs of emergency responders, and develop behavioral computing approaches to mine noisy social media datasets.
Combining data mining with social media could lead to better and faster emergency response to such devastating crisis situations as earthquakes, floods, and hurricanes. Information sciences and technology assistant professor Hemant Purohit studies human behavior on the web and leads Mason’s Humanitarian and Social Informatics Lab. Purohit has a nearly $175,000 two-year NSF grant (starts in September and ends in August 2019).
Purohit will be working with the Fairfax Fire and Rescue Department and a social media working group for emergency services at the Department of Homeland Security Science & Technology Directorate, among other groups, to refine the required information needs of emergency responders, and accordingly, develop behavioral computing approaches to mine noisy social media datasets. Current practices are slow and limited to keyword searches and filtering irrelevant information from social media. The volume and questionable accuracy of the tweets also have been stumbling blocks to effectively using this unconventional information source.
Purohit and team will work with emergency response teams to identify meaningful intentional behavior categories that align with emergency response needs, followed by working with psychologists to further validate behavioral categories and develop categorization algorithms. They’ll develop and evaluate a tool that uses these categorization algorithms to highlight social media posts most helpful to emergency responders. The software will be made publically available through open source code and promoted to the emergency community and other interested parties. Plus, Mason students will be using the research results as part of MS AIT and Data Analytics Engineering courses.