Report Open Access
Luba Katz, PhD
Data science is an approach to scientific inquiry that uses computational, statistical, mathematical, and domain-specific tools to extract knowledge from a large volume of data. Its potential for discovery is increasingly recognized by the academic community, as illustrated by the growing number of new degree programs, research centers, initiatives, and departments in the United States. A recent review of websites for 116 research universities revealed that more than 80% included at least one data science offering (Exhibit 1).
This report describes the mission, organization, and activities of data science centers, initiatives, and departments at 20 of these universities (Exhibit 2). The initial set of institutions to include was recommended by the Sloan and Moore Foundations, which funded three of the data science centers through its Moore-Sloan Data Science Environments (MSDSE) program.1 The sample was expanded with suggestions from the participants. Because of this ad hoc sampling strategy and some level of nonresponse, this report is neither representative nor inclusive of all efforts in data science, but rather is an attempt to capture a range of models being explored by universities.
For 17 of 20 centers, the information presented is based on three sources: (a) one hour telephone interviews with the leadership of the centers conducted between December 2016 and June 2018, (b) review of the center websites and materials provided to us, and (c) a short survey of participants in the Data Science Leadership Summit held in October 2018. For the remaining three centers, at the New York University, the University of California Berkeley, and the University of Washington, we collected extensive additional data as an external evaluator of the MSDSE program.
In the next section, we describe the organizational models, programs, and activities at 20 universities, illustrated with a few specific examples. Following, we include short profiles of each site, which contain additional information. We caution the reader that in our experience data science entities are rapidly evolving, and therefore some data presented in this report may be out of date.