Trio awarded “Best in Show” at DataFest

By Talia Bank ’23

“Data is more about how you tell a story than the analysis itself. Without the context, it’s very hard to tell a good story with the data.”
—Thu Dang ’23

In late March, Elm City Stories—a video game designed to educate 11-14 year-olds on risk reduction, substance abuse, and sexual health—saw a sharp uptick in engagement from college students across the country. But it wasn’t a sudden interest in avoiding risky behaviors that drove these students to the game.

Rather, they were participants in the 2022 DataFest competition, a nationally coordinated annual event where teams of two to five students work together over a weekend to analyze, visualize, and interpret a complex dataset. This year’s competitors, including sixty-five Macalester students mentored by faculty members, doubled down on a dataset about the use and effectiveness of Elm City Stories, tuning in from over seventy-five colleges and universities in the US, Canada, and Germany. 

Thu Dang ’23, Duc Ngo ’22, and Bea Bautista ’24, all participants of previous DataFests, were awarded the “Best in Show” prize for their entry, titled “How can we predict who takes Elm City Stories seriously?” 

“I feel like this team has been the dream team because there’s a very good balance between technical and storytelling skills,” Dang says. “We broke down the big question into smaller manageable tasks and divided them among ourselves, including playing the game to understand the context, analyzing the data, and preparing the slides. Data is more about how you tell a story than the analysis itself. Without the context, it’s very hard to tell a good story with the data.”

Given the game’s purpose as an educational resource and risk prevention tool, the team decided to focus on whether its users’ behavior was influenced after playing it.

“Someone who just wants to get it done because they’re given Elm City Stories as a school assignment could just click through everything and not really care,” Bautista says. “It emerged in our conversations that it would be interesting to look at who’s taking the game seriously versus who’s not.”

Over three days, the team cleaned up and organized the massive dataset and isolated variables such as how much time adolescents spent playing the game and how they scored on follow-up surveys gauging the likelihood of drug refusal and behavior in other risky situations. They found a correlation between time spent playing the game and scores on the periodic surveys; the players who spent the least amount of time tended to fluctuate most on their survey scores. These results led the team to identify players who spend more time as those who take Elm City Stories more seriously, with their counterparts who spent less time taking the game less seriously.

For Dang, her team’s accomplishment had a personal takeaway as well. “All the hard work pays off,” she says. “If I can come from no background in data as a first-year to winning this prize, I want to empower all the girls out there: they can do it, too. Don’t be intimidated. If you want it, you can do it.”