Interpreting Big Data for Smarter, Safer Communities
Fire departments across the country rely on after-action reporting by firefighters to keep records of events.
They make use of these records when crafting future policy decisions, so it is extremely important that these records be as accurate, as specific and as reliable as possible. Dr. Aimee Roundtree, Associate Dean for Research and professor of English in the Texas State College of Liberal Arts, is using text mining and smart-city ideas to help firefighters keep better records of their work.
If the wrong code is chosen, or if the narrative is not accurate and concise, it can make it difficult to track and identify useful trends and plan ahead for potential emergencies.
“All of these decisions that affect people’s lives hinge on whether or not the data are accurate, and whether or not the data are accurate hinges upon whether or not they have been reported accurately,” Dr. Roundtree said. There are often inconsistencies in the quality of reporting by first responders. This is further complicated by the fact that there are over 1,000 different codes in use and some codes, like “false alarm,” are used to describe a wide variety of possible scenarios. If the wrong code is chosen, or if the narrative is not accurate and concise, it can make it difficult to track and identify useful trends and plan ahead for potential emergencies.
Dr. Roundtree uses natural language processing techniques, along with interviews and surveys of first responders, to look for patterns and features of effective reporting. She then provides resources to the departments like reporting templates, decision aids and reporting workshops that help first responders select the codes they use and write quality reports.
The project grew out of a National Science Foundation grant application aimed at using big data to improve first responder services in urban areas. Roundtree’s research is currently backed by a State Farm Insurance community grant and is being tested in San Marcos and College Station.
Accurate as of September 2018