To work toward a leadership position at the confluence of these areas of responsible data science, the Data Science Task Force recommends four overarching goals and a set of actions for each goal. The actions are listed from relatively small and immediate ones to larger and longer-term transformation. These recommended actions are described in greater detail in the final report of the task force.
Goal 1
Create shared understanding. Increase the reputation, visibility, and awareness of responsible data science within and outside the Pitt community. Create a shared and unified understanding of data science and of its importance across disciplines.
Action 1: Establish a group of “data science liaisons” to ensure diverse representation of faculty, staff, students, postdoctoral fellows, and alumni to form an initial community to seed, welcome, nourish, and mentor the growth of a larger, inclusive community of individuals using, critiquing, governing, and regulating data science, as well as individuals who have curiosity, new interest, or need for expertise with data, but are finding impediments to doing so. (short term)
Action 2: Establish a regular University-wide/Provost “Distinguished Data Scientist Lecture Series” to invite distinguished visitors to advise and speak on uses, methods, ethics, laws, and critique of data and data methods. (short term)
Action 3: Unify, market, and communicate “Data Science Success and Opportunity” that reinforces a message of responsible, use-driven data science. (short to medium term)
Action 4: Create and continuously update a “Data@Pitt” online web hub to aggregate and disseminate opportunities, success stories, events, activities, education pathways, and initiatives related to data science. (short to medium term)
Goal 2
Require fluency and knowledge. Require every undergraduate student to acquire a basic understanding of data and data methods, including considerations of responsibility, as part of their learning at Pitt.
Action 5: Mandate that every school with an undergraduate program develop inclusive curriculum, coupled to practical experience with actual datasets, questions, methods and tools, that provides all undergraduates with preparation in data concepts and skills. (medium term)
Goal 3
Catalyze skill acquisition. Create, support, and incentivize inclusive, flexible undergraduate and graduate educational programs and shared educational resources to offer training in data science – in context of a broad variety of domains – to students, postdocs, staff, and faculty.
Action 6: Identify gaps in existing curriculum, develop a set of shared educational resources for these gaps, and provide a central repository of curricular materials at both the undergraduate and graduate levels. (short to medium term)
Action 7: Establish and/or charge an organizational entity to coordinate training and education, assessment, development of curriculum, collecting and disseminating project opportunities, courses and course content materials. (medium to long term)
Goal 4
Coordinate strategy and action. Implement a structure that (i) knits together, in a visible, accessible, and central place, people and practices in data science; (ii) serves as an evolving source of knowledge in incentivizing, developing and applying responsible data science to overcome diverse, challenging problems, including ethics, policy, and legal aspects; and (iii) animates extraordinary ambitions and success in collaborations transcending disciplinary and community limitations.
Action 8: Establish a dedicated, full-time position and charge a leader with a mandate to advance and coordinate data science for Pitt. (short to medium term)
Action 9: Use the Pitt Momentum or other funding mechanism to encourage, initiate, and support action on the highest impact actions in this report and work toward the development of a polished concept, with pilot implementation, that can be a springboard for a major gift. (short term)
Action 10: Create and support an institutional structure – a “coordination tower” – to coordinate and incentivize existing and emerging elements (layers) of data science. (long term)