Columbia’s Data Science Day Highlights Transformative Possibilities

Apr 13 2017 | By Jesse Adams | Photos by Timothy Lee Photographers

Big data’s astounding possibilities for reshaping health care, business, government, and other aspects of our lives took center stage at Columbia University’s Data Science Day, a gathering of more than 600 engineers, researchers, and industry leaders at the Data Science Institute. In addition to presentations and predictions from a variety of practitioners who approach the use of data from very different perspectives, the event showcased demonstrations by students and faculty working on cutting edge research.

Students and faculty discuss their latest research during Data Science Day.

In a big-picture analysis of the opportunities and perils of data science, keynote speaker Alfred Spector, chief technology officer and head of engineering at investment management company Two Sigma, discussed how big data is challenging paradigms, from privacy to security, legal liability to the replicability of research studies. Data science tends to be better at working on complex but relatively straightforward tasks, such as maximizing revenue and clicks, he said, than it is at navigating murkier and more subjective challenges of human life.

“There are many benefits to data science but also inevitable concomitant risks and ethical questions,” Spector said. “Consider free will: is behavioral targeting good or bad? Humans have always tried to influence behavior, but something may be different at this scale of compulsion.”

Experts from across the university also offered insights into how data is being leveraged in divergent fields. Assistant Professor of Electrical Engineering Mingoo Seok and Associate Professor of Neuroscience Stefano Fusi talked about their work designing brain-inspired learning machines and how theoretical neuroscience is informing novel neuromorphic chips that are compact and low-power. A trio of Columbia University Medical Center faculty discussed new apps advancing the frontiers of patient-driven healthcare, including software that analyzes nurses’ notes and assesses risk factors for endometrial disease. In a panel on data-enabled sharing economy including faculty from the law and business schools, Assistant Professor of Civil Engineering and Engineering Mechanics Professor Xuan (Sharon) Di delved into her work developing shared mobility systems to carry more passengers while reducing traffic congestion.

Suman Jana, assistant professor of computer science, also spoke alongside faculty from Columbia’s Graduate School of Journalism and School of International Affairs about the implications of increasing connectivity and the challenges it creates for data security.

“Connecting almost everything to the internet creates new and exciting applications, but also a many-fold increase in the attack surface and the need to securely and automatically update and patch all software running,” Jana said. “We need an interdisciplinary approach incorporating cybersecurity, privacy, law, policy, and other fields.”

The day concluded with students and faculty sharing an array of demonstrations (see photo slideshow below) and data-related research on topics ranging from analyzing crop yields to preparing for climate change to treating diabetes. Several groups affiliated with the Data Science Institute, part of Columbia Engineering, presented their latest research, including the Center for Smart Cities, the Center for Health Analytics, and the Center for Data, Media, and Society.

“Data science is revolutionary in solving problems in every field,” said Kathleen McKeown, director of the Data Science Institute and the Henry and Gertrude Rothschild Professor of Computer Science. “To keep up, DSI has recently added over 30 new faculty, both certification and masters programs, and an intro to data science for non-scientists.”

Stay up-to-date with the Columbia Engineering newsletter

* indicates required