Higher education is in a time of upheaval so I¡¯ll hazard a guess that relatively few of us who work at universities, even long-time administrators, have much experience when it comes to establishing a new school. For all the new labs, institutes and centres that universities propagate and nurture, the birth of an entirely new school is a comparatively rare event.

Our recent efforts to create the School of Data Science at the University of Virginia ¨C with a mission to radically transform the way we think, teach and work ¨C have sparked a significant amount of curiosity among our peers at other institutions. I won¡¯t pretend that this interest is solely rooted in the school itself.?

Rather, it seems to be a function of the infrequency with which new university schools are formed and the fact that we are, as far as we can tell, the first independent, dedicated school of data science in the country.?Thus, in the spirit of the openness and collaboration we espouse in our scholarship and practice, and to help serve as a template for the future, here are five things we learned along the way.

1. Don¡¯t start from scratch

We started first as the Data Science Institute where we were already operating similarly to a school: we awarded degrees (masters in data science), hired non-tenure track faculty and contributed to leadership decisions that affected the entire university.?

ÁñÁ«ÊÓƵ

This level of development and profile enabled us to form durable partnerships ¨C with other schools at the university, with the community, with corporations and government agencies and with non-profits ¨C that will continue and grow with the school.?

?2. Timing?is important

The timing of our development was fortuitous in that it coincided with new leadership at UVA. With new leadership came the opportunity to harness the new energy that comes with any such transition. We were able to capitalise on the new president¡¯s and provost¡¯s early recognition that establishing the university as a leader in data science represented an essential growth opportunity.

ÁñÁ«ÊÓƵ

Without the constraints of precedent or deeply rooted administrative structures, we set out to form a school that we believe embodies the fundamental tenets of data science ¨C responsibility, collaboration, openness, diversity and inclusivity ¨C through a process that similarly reflects those tenets and the iterative, inventive nature of the discipline.

3. Being new is exciting, and a little scary

As an emerging discipline, data science is overflowing with opportunity. It is not entirely hyperbolic to say that research opportunities in the discipline are limitless and the number of interested students far exceeds what any single school could hope to enrol. The demand for data-science skills, both from industry and students, is likely unmatched by any other discipline in UVA¡¯s history.

Given this demand, and the boundless possibilities for innovation, convincing the university community and the public of the value of data science represented a relatively straightforward endeavour. But communicating the value of creating a school dedicated to the instruction and application of data science was more involved, primarily because any new entity can be perceived as a drain on a finite pool of resources. And, in the process, it represents a threat to established schools and disciplines that rely on those resources.

Forming truly collaborative partnerships across schools and units was essential to our success.

ÁñÁ«ÊÓƵ

4. You will need help along the way

We devoted a tremendous amount of time to engagement with the deans of the other schools, illustrating how the necessarily interdisciplinary and collaborative nature of data science ¨C as we conceive of it and practice it ¨C serves to emphasise and amplify the work of the entire institution.?

Throughout our journey, we should have devoted an equal amount of time to engagement with faculty and participated more fully in the faculty senate process for?. Since the school¡¯s approval we have increased and broadened our efforts to collaborate not just with leadership, but with faculty and staff across the university.?

In our case-making, we emphasised ¨C and continue to reinforce ¨C that our work, our resources and our expertise, are not strictly ours. They are applicable equally across all disciplines.?

5. Share the love

We don¡¯t own anything. We can only be successful if we join other schools of the university to engage in collaborative, open, shared research endeavours. And this research is in service to the university, local community, state, nation and the globe; part of an overall effort to use data science as a means to seek the truth and maximise societal benefit. It will be our ongoing goal as a newly formed school to widely transmit this mission and imperative.

ÁñÁ«ÊÓƵ

As we continue on our journey, hopefully setting an example for other institutions to follow, we will continue to share what we learn in the spirit of openness, collaboration and discovery, and listen to all that you offer in return.

Phil Bourne is the dean of the newly formed University of Virginia School of Data Science and previously served as the associate director for data science at the National Institutes of Health.?

ÁñÁ«ÊÓƵ