Data Science Institute Director Reflects on First Year

In May, the Data Science Institute graduated 44 students with the University of Virginia’s first master’s degrees in data science, capping a year in which it also obtained powerful new computing capabilities for the University, partnered with companies on its student capstone projects, expanded its staff, played a key role in attracting “Big Data” research faculty, offered a distinguished lecture series, completed an executive education series and watched students flourish in the viable and growing field of data science.

UVA Today got the details from institute director Don Brown, the William Stansfield Calcott Professor of Systems and Information Engineering in U.Va.’s School of Engineering and Applied Science.

Q. The University has hired several new faculty members in the past year, some with “Big Data” expertise. Is DSI part of that effort?

A. Yes, we helped hire several key new faculty members, such as Karen Kafadar, an accomplished statistician in forensic science, who played a major role in making U.Va. a partner in the new National Institute of Science and Technology Center for Excellence in Forensic Science. She also serves as chair of the Department of Statistics, and uses big data methods in her research. We also helped bring in Denis Nekipelov, a big data economist and computer scientist who helped arrange and invite major speakers for the DSI Distinguished Lecture Series. Both Denis and Karen are serving on the DSI’s advisory board. And Joan Bienvenue, director of the Applied Research Institute, also fairly new with the University, has brought in a large big data-related multidisciplinary grant from the Navy that is bringing ongoing research opportunities to our faculty.

In all, 11 new faculty members have come to the University in various departments of the College of Arts & Sciences, in the Engineering School, in the McIntire School of Commerce, the Darden School of Business and elsewhere, bringing big data skills to help transform research at our University. And we’re in conversation with many other folks.

Q. How does DSI attract them?

A. We draw them because we are creating a collaborative environment in which people can do a range of data science projects, often in concert with others who are doing related or even not obviously related work. This is very appealing to researchers who are competing nationally for multidisciplinary grants. We also are offering ever-improving infrastructure, such as greater bandwidth, and our new high-performance computing cluster, Rivanna. The goal is to raise the level of research and education in data science across Grounds so U.Va. is a recognized leader in cutting-edge, data- and computationally intensive research.

We also have purchased a new $2.25 million secure cluster in conjunction with the U.Va. Health System, that has a virtual machine environment for computational flexibility and almost 2.75 petabytes of storage for handling large amounts of diverse data. This will enable authorized users to use protected health data for research to better understand disease trends and patterns while ensuring privacy. The same system can also be used by researchers in other domains, such as education, to find patterns and associations in large collections of student records. This big data acquisition will also help attract leading data science researchers to U.Va., and certainly lead to important new published papers and innovative research proposals. It will allow us to conduct important research with national implications in many domains while still protecting the privacy of the data.

Q. One of the hallmarks of the Master of Data Science degree is that students conduct major research as a capstone to all they’ve learned in their coursework. You’ve also obtained sponsorship for some of these projects. Please tell us about that.

A. Our master’s degree program in data science requires that all students participate in a capstone project. These are projects for real clients in industry and government. The students are organized in teams and each team must apply the principles and techniques of data science to address an important problem for their client.

In this first year, we had 15 sponsored projects from a wonderfully diverse set of domains that included health care, education, finance, social media, safety and security, energy and emergency response.

Leidos sponsored three projects in the areas of energy, health and cybersecurity. The one on energy won a best paper award in the business and government track in the IEEE Systems and Information Engineering Design Symposium (SIEDS 2015). In its work, this team developed new approaches to predicting energy use for individual buildings at U.Va.

Another team sponsored by Northrop Grumman won a SIEDS 2015 best paper award for its work developing models to predict severe sepsis in adult intensive care unit patients.

While not everyone can be an award winner, I was very happy with all of the projects. This year’s results showed the power of these real projects as the culminating event to cement the learning objectives of our intensive 11-month program.

Next year, we will make the capstone project a two-semester experience rather than only a spring project. This will allow a richer, more in-depth look into the many fascinating patterns in the variety of data sets available to our students through these projects.

Q. You’ve also had some new hires within DSI. Who are they?

A. It’s a new program, but our team includes myself; Arlyn Burgess, our executive coordinator; Michele Claibourn, the research data services lead at the library and the DSI’s associate director for infrastructure and service; statistics professor Jeffrey Holt, our graduate program director; and Jim Harrison, the director of the division of biomedical informatics and the DSI’s associate director for health data.

We also soon will hire a new assistant director of career services, who will share an appointment with University Career Services. Our goal is to help students connect with employers building and working within the interdisciplinary field of data science and ultimately be able to place close to 100 percent of our graduates in data science careers. There is tremendous opportunity for graduates in this growing field, and we intend to put extra focus on identifying employment opportunities.

Q. Have many of the new graduates have found work in the field so far?

A. Of those who responded to our employment survey, two-thirds, or 25, have jobs at more than 20 different companies. Of the remaining third, most just started looking for jobs after graduation. Our graduates are finding employment from Virginia to New York, Arizona and beyond. They are working in finance, insurance, banking, a variety of scientific fields, media, hospitality and hotels, and government. Our most outstanding graduate, who maintained a 4.0 grade-point average, has taken a job with a company right here in Charlottesville.

Q. What sets DSI apart from similar organizations at other universities?

A. We are unique in that our programs, including our master’s, are designed to be interdisciplinary, bringing together experts and students from across Grounds to share, collaborate and find connections that otherwise might not be obvious. Companies and funding agencies like this feature, and as we get the word out about what we’re doing here – that it’s not just in a single discipline – we will continue to prove to be highly competitive at placing graduates, winning grants and conducting needed research of importance to the state and nation. And we’ve got about $20 million in funding from gifts and the University with which to do these things.

Q. So how do you rate the institute’s first year?

A. Overall it’s been a very good year. We graduated all of our students, and some have chosen to stay on for more academic training and additional experience. This indicates that there may be strong demand for an extended program and possibly further refinement of our curriculum.

Already we are making modifications for the next year. We know that the students are pleased with the program and the strong interaction they’ve had with other students on their range of projects. Over the next year, we will be offering new courses leading to a minor in data science that will be open to undergraduate students from all schools and majors at U.Va. We also are looking at possibly working with the School of Continuing and Professional Studies to offer online courses leading to certificates in data science.

Stay tuned; this institute will keep getting better.

Media Contact

Fariss Samarrai

University News Associate Office of University Communications