DataFest 2022 at Duke

DataFest 2022 at Duke, held April 1 - 3, was one of over forty ASA DataFest events that took place March - May 2022. DataFest is a data analysis competition sponsored by the American Statistical Association, where students work in teams to analyze a surprise and complex data set over a weekend. Students from Duke University and Duke Kushan University participated in this year’s event, the first held in person since DataFest 2019. 

The surprise data came from Elm City Stories, an educational video game developed by researchers at the Play2Prevent Lab in the Yale School of Medicine aimed at middle and high school students. The data included logs from over 150 players capturing various actions players took as they played the game. Because the game’s primary objective is to promote healthy behaviors among adolescents, there was also survey data that measured students’ ability to make healthy decisions in different example scenarios. Participants were tasked with using these data to derive insights about game play patterns and the effectiveness of the game in helping adolescents make responsible decisions. 

Each team presented their analysis on Sunday afternoon to a panel of judges in statistics, data science, and related fields in academia and industry. The following projects were recognized as this year’s winners:

• Best Visualizations: A predictive tool to identity at risk adolescents by Jenna Boguslavsky, Rebecca Combs, Emily Du, Helene Gu, Athena Yao

The judges were impressed by this team’s effective use of the entire collection of visualizations to effectively communicate their results through multiple channels.

• Best Insight: Reordering minigames with personalized Recommendation System by Jianbo Bin, Yuge Jin, Chengkai Peng, Jingnan Wang, Yijie Wang

The judges were impressed that this team recognized a real problem of people not finishing the game and created a recommender system to do something actionable to address it.

• Best Use of Statistical Analysis: Investigation into Elm City Stories’ MiniGame Design by Bridge Hu, Andy Liu, Jacqueline Ma, Andy Wulu, Cara Yang

The judges were impressed that this team carefully played the game in order to decide to focus on minigames. They did a great job with the data pre-processing, and the judges liked their careful analysis of the data.

• Judges’ Pick: Evaluating Character Differences in the Top and Bottom Performing Students with a Custom Score by Joseph Ilagan, Aryan Mathur, Lina Nayvelt

This award is for a project recognized by the judges for outstanding achievement outside of the core award categories.

The judges were impressed by this team’s great presentation and that they focused on personality and psychology. The team outlined their objectives clearly, the real-world application, and discussion.

• Honorable Mention: Exploring effectiveness of the game by Xiaoyin Kang and Ziying Ye

The judges were impressed by the fact this was the only team who made an important distinction of the data in terms of users vs. the avatars of specific players playing multiple times under different identities.

The winning teams’ presentations and writeups are available on the DataFest website.

The Duke’s Academic Media Production team documented the event and interviewed participants throughout the weekend for the Duke Center for Computational Thinking. Click here to view the video and hear from students about their experience at different points during the event.