The eScience Institute recently spoke with six Student Fellows from the Data Science for Social Good program, now in its 10th year at the University of Washington’s Seattle campus. In-depth project descriptions are available on the UW DSSG projects page.

The Seattle summer is in full effect. People are taking advantage of the extended daylight hours with throngs of festival-goers, outdoor enthusiasts, and urban adventure seekers making the most of the Pacific Northwest at this time of year.

Participants in the UW Data Science for Social Good program have been striking a balance between summer fun and the intensity of their summer internship. “I did the Skyline Trail last weekend, and it was one of the most beautiful hikes I’ve ever done,” said Apoorva Sheera, a graduate student at the University of Washington (UW), speaking of Mt. Rainier looming in the distance on any given day (depending on the weather).

“The opportunity to come to Seattle was definitely a thing,” started Sakshi Chavan, a graduate student visiting from the University of Wisconsin-Madison. “It’s really different from our Madison campus. I feel like UW is really bound by nature, which I love.”

The city is in the middle of a multi-week heat wave. In all directions from campus, there are steel-blue skies and long, hot afternoons. Typically bustling with students, the campus is quieter during this time of year between the spring and fall semesters. The iconic Drumheller Fountain is switched off and undergoing maintenance. UW is in summer mode. At odds with the slower pace, the southern edge of campus hums with the ebb and flow of cyclists along the Burke-Gilman Trail. Adjacent to the trail is the Physics and Astronomy Tower, with the 6th floor home to the University of Washington’s Data Science for Social Good (UW DSSG) program.

From a pool of over 100 applicants, 16 Student Fellows are selected for the annual 10-week summer program. Working in teams of four and with the guidance of seasoned mentors, Fellows engage in projects centered around specific societal challenges. Project domains vary from year to year, with the 2024 program selecting project applications from campus and external organizations covering transit equity and homelessness counts in the greater Seattle region, water reuse metrics for city governments across the U.S., and an analysis of bias in crowd flow models.

Supported through an onboarding process that emphasizes team bonding, ethics and human-centered design, as well as best practices in best practices in reproducibility, and version control, the experience is immersive and quite unlike typical graduate and doctorate-level programs. Unsiloed and interdisciplinary by design, the heart of the program is a vision for data science that includes participants from domains as disparate as ecology, psychology, and law.

“I think one thing that is exciting is the interdisciplinary touch. We have people from different academic backgrounds who come together and give their perspectives on a project, and our discussions go way beyond our time for the meetings because we have so many different opinions going on in a meeting,” said Manurag Khullar, a graduate student from the University of Pennsylvania.

Grind Time

UW DSSG is analogous to a marathon with peaks and valleys that teams push through during the span of the summer. It is week six of the program, and Student Fellows have passed the halfway point. It is almost the beginning of ‘grind time,’ as staff in the program like to say.

“Comparing this experience with the previous team collaborations I’ve been a part of, there are no free riders in [UW] DSSG. Here, people are more self-motivated, with everyone willing to contribute something to the project,” mentions Jiaqi (Kiki) He, a graduate student at UW.

“The most surprising thing for me has been the emphasis on writing good, clean code that can be put into open-source [projects]. This has not been my experience with research in general. I’m really happy to be involved with something that is contributing to reproducible data science,” continues Sheera. “Working in a team has been more forgiving in this scenario because it’s an academic environment; the focus is on learning rather than the output.”

To create the right mix of self-directed project work, stakeholder engagement opportunities, and mentorship, teams have dedicated support from multiple perspectives, including Project Leads deeply conversant with the problem space, Data Scientists from UW’s eScience Institute, and guest seminars covering a broad range of topics from GitHub best practices to career advice panels featuring government, academic, and non-profit professionals.

“[UW] DSSG might be the first time most Fellows get exposed to formalized team-based science, and the Data Scientist is their primary mentor and teacher in that learning process. Even though most science is team-based, graduate degrees are largely individual affairs. We Data Scientists have spent years learning how to get up to speed on collaborative projects and implementing processes to aid in teamwork. [UW] DSSG provides a safe environment in which Fellows can learn and practice the soft and technical skills required of team-based science, and the Data Scientist is their guide during that time,” said Curtis Atkisson, an Anthropologist, Research Scientist with the eScience Institute and lead Data Scientist for the Water Reuse project.

The environment cultivated by UW DSSG is a recurring theme in discussions with Student Fellows. Terms such as “open-minded,” “amazing,” and “knowledgeable” are frequently ascribed to Project Leads and Data Scientists within the program.

When asked about the Project Leads for the Crowd Flow team, Chavan and Khullar are both effusive in their admiration. “She’s just so passionate about the project. The way her brain works is really fascinating to me. She’s always thinking about something and connecting the dots. It definitely makes the project more exciting,” Chavan said, referring to Afra Mashhadi, Assistant Professor of Computer Software and Systems at the University of Washington’s Bothell campus.

Similar praise is leveled at Ekin Urguel, Ph.D. Candidate with the College of Engineering at the University of Washington, and Co-Project Lead on the Crowd Flow project. “Same with Ekin. Super supportive during the process and at project discussions,” said Khullar.

“Having experienced the Indian educational system, some people might be hesitant to ask questions,” says Himanshu Naidu, a graduate student at UW. ” There can be a tendency to avoid asking what might be perceived as ‘stupid’ questions. In [UW] DSSG, however, the environment encourages and welcomes all questions.

“Pretty good snacks,” adds UW DSSG Fellow Ishan Saksena with a knowing grin when asked about the data studio environment. A visiting graduate student from the University of Michigan, Ann Arbor, Saksena has found the program to be particularly instructive for honing his skills.

“The tutorials and workshops are very interactive. Guidance from Naomi really encouraged me to actually change my coding practices,” mentions Saksena, referring to UW DSSG’s resident Data Scientist-at-Large Naomi Alterman, who mentors teams throughout the 10-week program with ongoing check-ins to gently guide and provide moments of reflection for the 16 participants.

“I’d say the setup is really nice. It doesn’t feel like your traditional classroom. I love that. Lots of natural light makes it feel balanced and you can relax and get work done,” offered Chavan.

With four weeks remaining in the program, the data studio is a hive of activity. Preparation for stakeholder engagement meetings, weekly presentations on team progress, deep dive working sessions, and emergent code challenges have the participants vacillating between focused concentration, moments of doubt, and exuberant exaltation at a problem solved. “All part of the process,” according to UW DSSG Data Scientists who’ve seen the same patterns play out each year within teams.

With the benefit of hindsight, Chavan reflects on the program midpoint: “I think I would tell myself that it’s going to be okay because I was quite worried in the beginning as to how do I fit in this space a little bit. Maybe a little bit of imposter syndrome. There are really great people who are super supportive, so you don’t need to be that worried going into it. It’s a beautiful city with beautiful people.”

Please visit the UW Data Science for Social Good website to learn more about this year’s projects and program participants. All are welcome to attend the team’s final presentations, which will be held on Wednesday, August 14th, 2024 at 1:30 pm.