Academic Data Science: from individuals to institutions

Although data-driven research is already accelerating scientific discovery, substantial systemic challenges still exist in academia that impact both individual researchers and institutional decision-making. These challenges need to be overcome for academia to fully realize the promise of the new data era. Toward that end, working in partnership with one another and with the Gordon and Betty Moore Foundation and the Alfred P. Sloan Foundation, three universities (the University of California Berkeley, New York University, and the University of Washington) have been attempting to create supportive environments for researchers using and developing data-intensive practices. Established in 2013, and known as the Moore-Sloan Data Science Environments (MSDSE), this collaboration was structured through a set of working groups on cross-cutting topics viewed as critical to advancing data science in academia: career paths and incentives, software development, education, reproducibility and open science, reflexive and reflective ethnography, and the role of physical space in collaboration. As the MSDSE grants approach sunset in 2020, the Academic Data Science Alliance (ADSA) has formed to expand the community and build a network across the U.S. and beyond to share lessons from the MSDSEs and from other higher education institutes experimenting with the integration of data science in academia. This talk will cover some of the efforts and activities of the MSDSEs and ADSA, emphasizing best practices and lessons learned that have emerged from six years of collaborative institutional experimentation, from cross-domain workshops and project incubators to the challenges of creating (and filling) new staff data scientist positions outside of any one particular lab or discipline.