More than 700 participants from around the world participated in the Data Science Day 2020 sneak preview on March 31. The virtual event was a prelude to the Data Science Institute (DSI) at Columbia University’s annual flagship event.

Data Science Day 2020 is rescheduled for September 14 due to the coronavirus pandemic.

“We wanted to offer you a virtual preview of what you’ll see in September,” said Jeannette M. Wing, DSI’s Avanessians Director, during her live-streamed opening remarks. “We hope it’s an inspiring experience for all during such challenging times.”

Wing’s opening address highlighted three research challenge areas—causal reasoning, trustworthy AI and ethics—that are driving and challenging the study of data science. She noted that the areas shouldn’t be seen as a definitive encapsulation of the field, but rather as a way of  “kickstarting a discussion of our broader research agenda.”

Participants were invited to explore 27 research posters and videos by Columbia students, professors, and researchers showcasing the breadth and scope of the university’s most advanced and innovative data-driven research and interact virtually with the research teams.

The areas of research covered by the teams included health care and medicine, finance, environmental science, physical science, wireless technology, computing systems, marketing, material science, urban affairs, media studies, and social work.

Aviv Landau is a DSI postdoctoral research fellow who uses data science to enhance traditional social work. During his virtual presentation, Landau discussed his efforts to develop an AI system that can objectively detect child abuse and neglect. He cited difficulties for clinicians to assess accusations of child abuse and systemic bias against Black and Latinx parents.

“We are building a system that doesn’t have bias against certain communities and will allow all families to be assessed and treated equally,” he said.

Landau collaborates with two DSI members: Max Topaz, a professor of nursing who has pioneered the use of natural language processing to process the data collected by nurses, and Desmond Patton, a professor of social work known for his data-intensive studies of social media and gang violence. The team received a DSI Seed Fund grant to support their research. 

Another virtual presenter and DSI postdoctoral fellow, Ipek Ensari, focuses on chronic pain diseases that have no cure and therefore may benefit from self-management. She develops models to predict pain fluctuations from exercise as well as how much these fluctuations differ among patients.

“Endometriosis is a poorly treated disease and our study provides evidence that we might be able to intervene on this highly unpredictable symptom through exercise,” Ensari said.

She collaborates with DSI members Noemie Elhadad and Suzanne Bakken, two leaders in biomedical informatics. Their findings could revolutionize the way researchers and clinicians approach pain management for patients with chronic conditions.

Data Science Day 2020 will feature keynote addresses by Eric Schmidt, technical advisor to the board at Alphabet Inc., and Pat Bajari, chief economist and vice president of core AI for Amazon. The day will also feature lightning talks by Columbia’s leading data science researchers. Registration for the Sept. 14 event is open.

— Robert Florida