For those who lived and wrote the story of data science at the University of Virginia, it is about much more than just the history of a school. For many of them, the creation of a school was far from their minds when this story began.
What would transpire over the years — and ultimately result in the creation of UVA’s 12th school and the first of its kind in the nation — was the product of fortunate timing, extensive outreach and planning, committed champions, and more than a little luck.
In the narrative that follows, we’ll trace the history of these efforts and the impact they would have. Later, the early days of the Data Science Institute will be chronicled, including the development of the master’s program.
Finally, you’ll learn about the creation of the School of Data Science, its mission, and the people who brought it to life and embody its purpose, culminating in the opening of the School’s new home in 2024 at the entrance to the Emmet-Ivy corridor in Charlottesville.
The story begins near the turn of the century, then picks up rapid momentum in the early 2010s as the eyes of the media and Washington begin to focus on big data and efforts that were already underway at UVA gain strength.
What would ultimately emerge would be a School Without Walls, dedicated to interdisciplinary research, collaboration, and the practice of responsible data science for the common good. Run by a handful of staff in the early days of the Data Science Institute, the School of Data Science would become what it is today through the commitment, dedication, ambition, and joy brought to the work by more than 100 faculty and staff, as well as an ever-increasing group of students and alumni.
For them, the story of data science at UVA is still being written, and it is one that began with just the seeds of an idea.
When Rick Horwitz came to UVA in 1999, he was looking for a new start to his research career. He’d most recently been at the University of Illinois, where he established the Department of Cell Biology, but he was ready to move out of administration.
“It was time to come to a place where I was unknown,” he said, adding that he was ready to set “a new direction” for his research. It would not take long for him to get that opportunity.
Shortly after Horwitz arrived at UVA, the National Institutes of Health launched a major new effort aimed at promoting large-scale, collaborative science. UVA would receive what was known as a “glue grant,” with Horwitz the principal investigator. The grant would fund an international cell migration consortium to better understand the invasion of cells.
The program spanned 10 years, and the research it produced required extensive collaboration across multiple disciplines. It also involved looking at data from a wide variety of sources, which got Horwitz thinking as the work of the consortium wound down: “Could I do something similar to catalyze collaborative research at UVA?"
In 2011, Horwitz moved into a new position — Associate Vice President for Research and Bioscience Programs — with a mission from his new boss, Tom Skalak, the University’s Vice President for Research, that was not without ambition. Skalak challenged Horwitz to “seed creative ideas and stimulate/motivate people to great achievements that were not visible before.”
Thinking back to his experience with the NIH glue grant, Horwitz was drawn to the idea of collaborations across Grounds but in work that went beyond just the biosciences. After pitching his broad, still-developing idea to Skalak, and receiving his blessing to move forward, Horwitz went to work.
“I went out, and I interviewed research deans and faculty,” he said, describing his next steps. “I just found anyone who’s doing quantitative stuff or had data or doing neuroscience,” he said, asking them about their work, their aspirations, and their data.
“Do you have data sets you’re not analyzing,” he recalled asking colleagues. “And if so, why aren’t you doing it? What are the bottlenecks, and what are the opportunities?”
He estimates that he interviewed 50 people. A theme was emerging — data, modeling — but, Horwitz said, Skalak, while supportive, was skeptical that Horwitz’s listening tour was yielding something concrete and actionable.
“There was no beef,” Horwitz said of his efforts at this time. “There wasn’t even a bun yet.”
Don Brown wasn’t sold the first time he spoke to Rick Horwitz about a big data initiative.
“I think I was nice to him,” he said. “But I was skeptical that this was going to lead to anything because I had watched UVA, multiple times, attempt various research initiatives, which had not gone well.”
When he met with Horwitz, Brown was certainly speaking with significant institutional knowledge. A West Point graduate and U.S. Army officer, Brown arrived at the University in 1985 after receiving his Ph.D. from the University of Michigan. In 1996, he became chair of the Department of Systems Engineering, a position he would hold for 12 years.
Early in his tenure at UVA, he was part of a team that received a substantial grant from the U.S. Department of Defense, which would be used to form the Institute for Parallel Computation, a collaboration with colleagues in the Department of Computer Science.
“You could call that sort of an early stab at some of these questions of data science and big data,” Brown said.
Despite his misgivings, Brown continued to meet with Horwitz.
“The thing about Rick,” he said, “was that he gave me a good story about why he thought the timing was right for big data.”
Soon, Brown was on board and their meetings continued, with a core group of around 10 faculty members from different schools across the University discussing what a big data initiative at UVA might look like.
Teresa Sullivan was no stranger to big data by the early 2010s. She worked with Census data — perhaps the crown jewel of big data sets — while completing her doctoral dissertation in the early 1970s at the University of Chicago.
“It was very time consuming, clunky compared to what you do today,” she recalled of the tapes she pored over in order to extract what she needed. It was tedious work, but she developed a knack for it, even catching an error that the Census Bureau wound up correcting. Decades later she would write a short book about the controversies surrounding the 2020 Census.
The 1970 Census may have been Sullivan’s first experience with big data, but it would hardly be her last.
In 1985, Sullivan, then a sociology professor at the University of Texas, was having lunch with two law professors, who were bemoaning the fact that it was difficult to assess whether recently enacted bankruptcy reforms were effective. Sullivan had an idea — look at the data.
She took the two professors — one of whom was future U.S. Sen. Elizabeth Warren — to the federal courthouse in San Antonio where they pulled a random sample of 150 bankruptcy cases and then coded them, a process, Sullivan recalled, that was much more cumbersome in the mid-1980s than it would be today.
Their hard work, though, yielded some compelling results, and they would eventually receive funding from the National Science Foundation and expand their study to other states. Their findings that bankruptcy laws were often not working as intended met some resistance. But the data ultimately won out.
“A lot of people said we were just wrong,” Sullivan recalled. “And there ensued a lot of controversy and some follow-up research. And our work passed from completely wrong to conventional wisdom in about five years.”
Sullivan would hold a variety of senior administrative roles at the University of Texas and the University of Texas System until 2006, when she was named provost and vice president for academic affairs at the University of Michigan.
After four years in Ann Arbor, Sullivan, in 2010, was tapped to become the University of Virginia’s eighth president. She would be the first female to hold the office and would serve for eight years. Her ascension to the top leadership post at UVA would prove fortuitous for data science at UVA.
Rick Horwitz and company would also enlist another important ally in John Simon, who took over as the University’s provost in September 2011. Momentum was building.
To gauge interest and bring together people with experience working with big data, plans were made to convene a summit focused on the concept that was developing.
The first Big Data Summit, a seminal event for data science at UVA, was held on May 9, 2012. The half-day gathering featured short presentations followed by panel discussions around the following themes: data producers; infrastructure issues and challenges; analytics and algorithms; and needs, opportunities, challenges, and next steps.
The overarching theme, though, was a clarion call to attendees to think creatively and without a preconceived notion of where this idea was headed: “Data, data everywhere without a plan in sight.”
By any measure, the summit was a success.
“It was just absolutely amazing,” said Horwitz. “It was beyond our expectations.”
“We were totally surprised,” said Don Brown of the interest and attendance, saying that they initially expected 60 to 80 faculty participants, not the 170 who attended from 32 departments.
“Far more than I anticipated,” Teresa Sullivan said.
The enthusiasm coming out of the first Big Data Summit, however, was marred, at least temporarily, by the forced resignation of Sullivan in June.
“It was devastating,” said Horwitz. After widespread protests, though, Sullivan was reinstated, and data science had its champion back in the president’s office.
In May 2013, a second Big Data Summit was held, with opening remarks from Brown, Sullivan, and Horwitz. The topics for this gathering showed the vast potential of data science to touch on a diverse range of disciplines and policies. Sessions focused on humanities, social sciences, and the arts; physical sciences and engineering; and biosciences and medicine.
There was no denying by the summer of 2013 that the time for big data at UVA had arrived; the next step was turning this vision into something concrete.
Sullivan asked Horwitz to present and promote the emerging concept around data science and a UVA institute at an early August retreat of UVA’s Board of Visitors. Their response was overwhelmingly positive, and they encouraged the group to move forward quickly with their plans.
Attention then turned to fundraising.
In 2012, a few moments occurred in quick succession as efforts at UVA to push big data to the forefront were ongoing.
A much larger, national spotlight began to shine on the issue. First, major national media outlets would take notice, including an in-depth look by The New York Times, which christened the era “The Age of Big Data.”
Then, the Obama administration put the weight of the White House behind the issue, unveiling the “Big Data Research and Development Initiative” in March 2012.
“In the same way that past Federal investments in information-technology R&D led to dramatic advances in supercomputing and the creation of the Internet, the initiative we are launching today promises to transform our ability to use Big Data for scientific discovery, environmental and biomedical research, education, and national security,” said John Holden, director of the White House Office of Science and Technology Policy, in a statement announcing the effort.
While these events helped galvanize outside interest, Sullivan notes that efforts at the University were well underway.
“We’ve been thinking along these lines already, so I don’t think we reacted particularly to the outside interest,” she said. “We had our own internal dynamic underway already.”
In fact, there were two internal dynamics. While Rick Horwitz and his group were developing their ideas, Sullivan was proceeding on a parallel path.
After arriving at UVA, she was invited to join the Business-Higher Education Forum, a group that convenes senior high education officials and CEOs to discuss trends. Not long after joining, she participated in a session on data science. She recalls CEOs describing how, increasingly, job candidates did not have the skills that were required, adding that they thought they would soon need to outsource positions due a lack of needed talent in the U.S. — a potentially “alarming” development, Sullivan recalled.
Around the same time, she met with Robert Groves, director of the U.S. Census Bureau, who told her that it was critically important that college graduates not just be data literate but also be able to work with very large datasets.
Sullivan would soon meet and begin talking with Horwitz: “We were both on the same track — we just didn’t know it,” Sullivan said.
Horwitz, Don Brown, and their colleagues no longer needed to worry about finding a champion in the president’s office for data science. “I was actually prepared to sell them,” Sullivan said.
Jaffray Woodriff’s passion for data began long before he ran a hedge fund. Like many, he was a baseball fan as a child, but unlike some, he was particularly enamored by the work of legendary, and pioneering, writer and statistician Bill James.
“I read as much as I could find of his writings, usually two or three times,” Woodriff said. “He was absolutely brilliant with data science and baseball statistics, yet even though he was publishing all of his work, he wasn’t even recognized as someone that knew anything at all about baseball.”
It wasn’t until decades later that Major League Baseball teams widely embraced the statistical tools James was championing in the 1970s. “This sort of highlights how slowly humanity adjusts to great ideas,” Woodriff said.
Woodriff studied finance at UVA, graduating in 1991. Even before enrolling he’d developed an interest in trading stocks and resolved that he would focus his studies on the goal of becoming a hedge fund manager. He attributes his success in trading stocks to his use of machine learning.
“Like Bill James was very early to apply data science to baseball statistics, I was very early to apply data science to trading stocks,” Woodriff said.
He would go on to found the Charlottesville-based firm Quantitative Investment Management in 2003, where he would continue to enjoy incredible success through his systematic use of machine learning in trading markets.
Another passion of Woodriff’s is squash, and he wanted UVA to field a team. But, Teresa Sullivan recalls, as with many aspects of higher education, there were resource limitations.
“I sat down with Jaffray and basically said, ‘Look, if this is what you want, you’re pretty much going to have to endow it because we’re just not going to have the resources in athletics to stretch it to another team,’” Sullivan says she told Woodriff.
Woodriff would end up supporting the establishment of men’s and women’s varsity squash teams and construction of the McArthur Squash Center.
“It’s certainly a world-class squash facility,” Woodriff said. “I’m very proud of that.”
“I think he was pleased not only that we were we able to do something new, but that we were able to do it on time, on budget, in a responsible way,” Sullivan said.
Woodriff’s support of new ventures at UVA was only just beginning. For years he had dreamed of a data science-focused school at the University.
“I had always wanted UVA to be the leader in data science,” he said. “And I couldn’t really figure out a good way to make that happen.”
An invitation to a dinner hosted by Sullivan would provide the opportunity he’d been waiting for.
Teresa Sullivan organized a dinner with potential donors, including Jaffray Woodriff, who asked Rick Horwitz about the plans to establish a Data Science Institute at UVA. Woodriff liked what he heard about the institute but had one concern: Who would run it?
“I’m worried that they’re going to put someone in charge who is really academically focused who doesn’t also have applied success in data science,” Woodriff said. In fact, he had someone in mind: Don Brown.
Woodriff had only met Brown briefly at a conference 12 years earlier but was immediately impressed and viewed him as possessing both the academic credentials and real-world experience — Brown had owned a successful data science consulting firm in addition to his role as a UVA faculty member — to make the institute a success.
As Woodriff sat at dinner, both excited about the prospect of his dream of a data science institute becoming a reality and concerned about who would be at the helm, he was finally informed that Brown, his ideal candidate, would be in charge.
“I think I stood up and punched the air with my fist in celebration that it was going to be exactly the person I thought was best,” Woodriff recalled years later, a moment Sullivan vividly remembered as well.
“Don was exactly the kind of person he had seen as the leader for this effort,” she said.
Woodriff would ultimately donate $10 million to support the institute — a gift that would prove to be just the beginning of his support of data science at UVA.
“With his first gift, it became legitimate,” Horwitz said of the impact of Woodriff’s donation. “We had money to start doing things.”
Other external supporters would also emerge. Brown recalls that another UVA alum, Scott Stephenson of Verisk Analytics, wrote to the School of Engineering promoting the idea of a program centered around big data and industry. While the Engineering School did not act on it, Brown caught wind of the letter and did follow up.
"I said, 'Scott, we want to do what you're talking about,'" Brown said. "It was like the melding of minds, we came together," Brown said. Stephenson would provide substantial support over the years and would go on to chair the advisory board for the School of Data Science.
The Data Science Institute was ready to begin its work.