Using 911 data from Arlington County Fire Department, SDAL's data scientists have identified factors that affect response time to structure fires—insights that can be used to improve the general safety of Arlington County residents and make the allocation of emergency resources more efficient.

fire truck driving to scene of 911 emergency

Problem

Fire departments are under strict pressure to respond to emergency situations efficiently. This pressure comes from many sources, including the taxpayers they service, as well as codes and standards organizations like the National Fire Protection Association (NFPA). The NFPA recommends professional fire stations maintain turnout times under 80 seconds, and travel times under 240 seconds. Turnout time is defined as the time interval that begins when a unit is dispatched, and ends when the unit leaves the station. Travel time is defined as the time interval beginning when the unit leaves the station, and ends when the first unit arrives at the scene.

Methods

Using five years of 911 data, SDAL researchers examined the effect of various predictors, such as hour of the day, station, year, month, and apparatus type, on turnout time and travel time. They also captured a spatial effect on travel time using a spatial Gaussian Process model.

By modeling turnout time and travel time separately, researchers are able to provide more detail about how Arlington fire stations can improve overall response times. For example, the results suggest that Station 106 is the slowest to turnout, yet among the fastest station to travel to incidents.

heat map showing Arlington County 911 Response Times

charts showing 95% confidence intervals on turnout time by apparatus and station

Impact

This research identifies regions of Arlington County which are vulnerable to slower response times, and equips Arlington County Fire Department with invaluable insight into their strengths, as well as opportunities to improve Arlington County residents’ general safety, and allocate resources more efficiently. Because the 911 data used in this analysis is uniform throughout most of the United States, this research could potentially be replicated in other jurisdictions to improve emergency response services across the country.

Back to top