Students Use Data Science to Solve Society’s Problems

Posted July 10, 2015

Fires will happen. But what if the Atlanta Fire Rescue Department could determine which buildings are at greatest risk of fire and prioritize inspections to focus on those sites?

The fire department is working to do just that thanks to a summer internship program sponsored by Georgia Tech and Oracle.

The city agency is one of four groups participating in this year’s Data Science for Social Good (DSSG), where 14 undergraduate and graduate students show nonprofits and government agencies how data can tackle social and societal problems.

Other projects include maintaining and improving Atlanta’s urban forest, increasing efficiency for a popular hotline system, and working with Georgia Tech’s WiFi data to understand mobility patterns.

“It’s important to get students involved in real-world projects,” said Christopher LeDantec, co-director of the program and an assistant professor of digital media in the School of Literature, Media and Communication. “We are building out sustainable learning opportunities that allow students and Georgia Tech to have a wider impact in the community. “

The four students working with Atlanta Fire Rescue shadowed inspectors to learn not only how they work, but also how they create and use data. The students identified hundreds of additional commercial properties at high risk for a fire that should be inspected with greater frequency, said Matt Hinds-Aldrich, a management analyst with Atlanta Fire Rescue working with the student team.

The program is “helping us make more informed and data-driven decisions on how best to allocate our limited resources as we balance protecting residents, visitors, and commerce in our community all while remaining fiscally responsible,” Hinds-Aldrich said.

While many public agencies and nonprofits collect data, staff may be overwhelmed by the volume or may lack the time and capabilities to interpret the information, said Bistra Dilkina, co-director of the DSSG program and an assistant professor in the School of Computational Science and Engineering in the College of Computing.

The program shows groups how data can work to their advantage, while exposing students to different career paths, Dilkina said.

“Nonprofits and conservationists need data scientists, too,” she said.

Georgia Tech received more than a dozen project proposals, but could only staff four. More than 100 students applied for the 10-week paid internship and 14 were selected.

Richard Huckaby, a rising third-year student in computer engineering, joined because he wanted to learn new skills and help people.

He’s part of the team working with the United Way of Greater Atlanta on their 211-hotline system, which is a searchable database of city services. Students will suggest ways to improve and reconstruct the existing call menu to better organize incoming calls and reduce average wait time.

“There’s more to data science than commercialization or helping large companies,” he said.

To conduct their work, students pulled data from multiple sources and had to go through several steps before it was ready for analysis.

The urban forest team working with the City of Atlanta and Trees Atlanta relied on tax parcels, thermal maps, city inventories, floodplain data, and other sources.

“Games and apps are fun, but we’re setting up a system the city and Trees Atlanta can build on in the future,” said Caroline Foster, who will start Georgia Tech’s master's program in Human-Computer Interaction next month.

Students from all four teams presented their findings and recommendations during a public demonstration last Thursday at Ponce City Market.

Originally featured in Georgia Tech News.