The Moore Sloan Data Science Summit Recap: Day 2

On cross-disciplinary data science research, & learning how to call out bullshit. Data scientists from NYU, University of Washington, & UC Berkeley gather to share their research

Day two in the Big Easy was jam packed as usual for our researchers at the Moore Sloan Data Science Summit.

Beginning with the Education panel, Richard Bonneau, Director of NYU’s Center for Data Science, Vicky Steeves (NYU), Magdalena Balazinska, Director of UW’s eScience Institute, Rosemary Gillespie (UC Berkeley), Michael Mahoney (UC Berkeley), and Anthony Suen (UC Berkeley) explained how the Moore Sloan grant has helped to develop various data science curricula and programs at the three institutions.

Mahoney in particular highlighted how they have been documenting the explosive interest in data science amongst UC Berkeley undergraduates by taking pictures of how many students attended the first day of the Foundations of Data Science course every year since 2015. Sure enough, the photos demonstrated how the crowd grew rapidly from 500 to over 1000 within years.

As interest in data science grows, however, so will the ways in which the methodology is applied to other fields — both in the physical sciences like biology, chemistry, and physics, and the social sciences like politics or cultural analysis.

As scientists partake in more cross-disciplinary collaborations, how can we ensure that researchers follow each discipline’s established norms while also pushing the field in new directions when working together?

This question was at the heart of Richard Bonneau (NYU) and Joshua Tucker’s (NYU) roundtable discussion, where they outlined the triumphs and challenges of working with their multi-disciplinary team at the Social Media and Political Participation Lab (SMaPP).

The rise of data science also means that we must continually think critically about the potential prejudices that influence data-driven results. Jevin West’s (UW) roundtable talk — colorfully titled “Calling Bullshit in the Age of Data Science Euphoria” — demonstrated how numerous computer vision and machine learning papers that tout themselves as objective are undercut by fishy methodologies, weak logic, and unconscious biases. West then opened up the floor for suggestions on how to improve the field. Some ideas included Associate Director for UW’s Data Intensive Research in Astrophysics and Cosmology (DIRAC) Institute and former NYU CDS Moore Sloan Fellow Daniela Huppenkothen’s suggestion that we need to instill an understanding that although data science is powerful, it also — just like every other discipline — has its limits.

The day ended with another round of lightning talks and the Career Working Group panel discussion consisting of Michael Laver, NYU Dean for Social Sciences, Ed Lazowska, Founding Director of UW’s eScience Institute, and Henry Brady, UC Berkeley Dean of the Goldman School of Public Policy. Presenting the results from last year’s survey about data science career paths, they found that, as data scientists, we not only have different definitions of what data science actually is, but that we also have different long-term goals. The survey revealed that the priority we all share, however, is maintaining our intellectual freedom.

The summit closed with a group dinner, drinks, and plenty of fun in New Orleans’s renowned French Quarter.

Several thousand thank-you’s are in order for Micaela Parker and Sarah Stone, Co-Executive Directors of UW’s eScience institute, Emily Mathis (NYU), Program Manager, Marsha Fenner (UC Berkeley), Program Manager, and Ali Ferguson (UC Berkeley), Communications Manager, for organizing such an exciting event.

See you all next year!

by Cherrie Kwok

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.