Veteran suicide is a pressing national issue in the United States. While veterans make up just under 9 percent of the U.S. population, they account for more than 18 percent of all suicides.
The Northern Virginia Technology Council sponsored a University Challenge “hackathon” to address the issue of veteran suicide prevention Feb. 28 at its Capital Data Summit, and #TeamUVA won top honors. The challenge was designed to present ways that big data analysis can be used to address a serious national challenge, and the summit showcased big data assets and capabilities in the region that can serve commercial and government interests, as well as society at large.
The team of five undergraduate students has been working for nearly two months with the UVA Data Science Institute to “hack” into the factors that can lead to so many veteran suicides, leveraging data science tools and techniques to analyze and propose solutions to the problem. Their sources included Department of Veterans Affairs statistics, demographics data from regions throughout the nation, death statistics and others.
During a recent internal competition at UVA the team presented its findings to four judges, each in the data science field, and won the internal competition, earning its place in the Capital Data Summit competition. The team members, and their competition at the summit, were judged on the basis of their choice of data, their analytical techniques, their understanding of and insight into the subject matter, and the quality of their presentations.
“This Northern Virginia Technology Council challenge led to a group of talented students looking at a very important social problem and increasing our understanding; it does not get any better than that,” said Phil Bourne, director of the UVA Data Science Institute, which is sponsoring the UVA team.
The team proposed ways in which social media and other data can be used to identify and locate at-risk veterans and offered solutions. Team leader Soukarya Ghosh, a first-year computer science major, heard about the UVA hackathon from a friend just a few hours before a planning session in January, and showed up “out of sheer curiosity.”
“The concept of looking at data which could cause tangible change in our communities is what really struck me and I was convinced to form a team to take on this challenge,” he said.
Ghosh’s team soon developed a program to analyze veteran tweets for patterns that might indicate behaviors and attitudes that could suggest depression and suicide.
“My group was able derive a number of correlations between veteran suicide rate and variables such as uninsured rate, urban/rural living location, and access to firearms,” he said. “Although these facets remain correlations, we plan on diving into deeper research to find root causes such as unhappiness caused by social isolation, poverty, etc.”
The team also further addressed shortcomings at the Department of Veteran’s Affairs, such as wasteful spending.
“There is no one right answer that will stop suicides once and for all; however, with small steps, like providing accessible healthcare to all our veterans and creating a strong community around them, can help them settle back into civilian life,” he said.
He added that the exercise is helping him collaborate effectively with his peers, and he plans to use the skills he’s developing to address additional large scale societal problems.
“Data science brings out opportunities for teams to come together and attack a problem that no one person could even begin to address alone,” said Arlyn Burgess, the DSI operations director who coordinated the internal competition. “Beyond their excellent technical abilities, the teams creatively found data sources and methods to address the way that we support our veterans when it comes to the difficult issue of suicide. These kinds of data-informed recommendations are going to change the world we live in, and we are excited to be part of it.”
Rahul Batra, a first-year computer science major on the team, signed on because of an interest in the interaction of technology and societal issues.
“Whether it is through something as simple as optimizing the way people interact with a system at a grocery store or through something broader like combatting bullying, technology always plays a large role in people’s behavior,” he said.
“When I heard that the DSI at UVA was offering this project I was thrilled because data analysis is a field that’s been completely changing the way companies operate, and it is interesting to see it now being taken to a policy level with relation to the government.”
Batra said he is happy to “tie together” science, technology and society concepts with computer science skills to work on policy objectives that could enact change.
“After collecting data, interpreting it, and running tests, at the end of the day we still needed to reason with it and create subjective policy recommendations,” he said. “This part of the project wasn’t something that we necessarily are taught in our classes, however it is equally important to our success as engineers to be able to communicate what our findings really mean.”
The merging of a range of skillsets by the team has impressed Don Brown, a chaired professor of Systems and Information Engineering, and one of the UVA competition judges.
“This talented team of students has given us new insights into the crisis of suicide among those who served our country,” he said. “As a veteran myself, I very much appreciate their work and am impressed by their technical accomplishments in a very short period of time.”