skip to main content

Social and Decision Analytics Laboratory, Arlington County team up

April 27, 2017

Sallie Keller at office desk
Sallie Keller

For statisticians like Sallie Keller, numbers equal opportunities.

“We look at data and think ‘collect, liberate, and repurpose,’” said Keller, professor and director of the Social and Decision Analytics Laboratory at the Biocomplexity Institute of Virginia Tech. “We continually question how we can use data in ways that benefit communities.”

Sometimes data even helps save lives, Keller said. 

Case in point: Keller’s lab and local officials in Arlington County, Virginia, joined forces to increase fire safety and improve emergency 911 response times.  

According to the National Fire Protection Association, the death rate per 100 reported home fires is more than twice as high in residences that lack an early detection system. To raise awareness about the importance of smoke detectors, Arlington County launched Operation FireSafe — it encourages residents to request installation of a free smoke alarm.

As part of this program, the county compiled data from thousands of home visits. Of nearly 2,000 successful home visits, 32 percent revealed the need for new smoke alarms, subsequently provided by the county.

Keller and her team at the Social and Decisions Analytics Laboratory — including economist Gizem Korkmaz and statisticians Josh Goldstein and Vicki Lancaster — used datasets from these visits and additional socio-economic data to develop a statistical model that predicted which geographical regions were most in need of fire detectors. Such information is helpful in streamlining fire officials’ door-to-door visits. 

Among variables used were the home’s value, age, location, number of bedrooms, and whether owned or rented. Socio-economic characteristics of the neighborhoods, such as income and education levels, were also included in the model.

“In addition to reducing the risk of fire-related injuries and fatalities, policymakers hope that a higher density of functioning smoke alarms will lead to improved emergency response times and less overall damage to property,” Keller said.

Improving allocation of emergency resources is another area where the researchers are lending support. Keller and her team, including economist and lab deputy director Stephanie Shipp and data scientist Aaron Schroeder, analyzed data from 911 dispatch centers throughout Arlington County in order to identify "loyal locations." These are key areas where people — generally disabled and/or elderly — used emergency medical services (EMS) resources most frequently.

“The county responded by hiring a nurse practitioner to help people who fell into this group understand how health issues could be managed without having to turn to 911 and when it was necessary to employ emergency services,” Keller said.

Cutting down on EMS calls allows the county to direct emergency response resources more efficiently to locations where they are really needed.

Keller said it has been rewarding to work with such forward-thinking officials as Arlington Deputy County Manager Jim Schwartz; Chief Information Officer and Director, Department of Technology Services Jack Belcher;  and Director of Emergency Management John Brown.

“Our work with Arlington County has laid the foundation for a broader collaboration with fire departments throughout Virginia and California to develop a new data science portal that will improve the health and safety of first responders,” Keller said.

Contact: