No Shortage of Data Degrees at Duke, More on Way
It looks like Duke University will launch yet another program that allows students to earn a master’s degree in data analysis, by the count of its advocates the eighth such effort on the Durham campus.
Supporters of the proposed master’s in “interdisciplinary data analysis” say it should complement rather than duplicate existing degree offerings, by helping extend number-crunching skills to “areas where data is only beginning to be recognized as important.”
And compared to data-degree offerings at other universities, “there’s an opportunity for Duke to present a distinct approach,” said Tom Nechyba, a Sanford School of Public Policy professor and director of Social Science Research Institute.
Nechyba offered that opinion as part of his pitch for approval from Duke’s Academic Council, a body that includes professors from all of Duke’s major academic units and that’s the closest thing it has to a faculty senate.
The council is one of the gatekeepers for new degree offerings, with Duke’s Board of Trustees having the final say when it comes to approving them.
Council deliberations on new degree proposals provide a window into the private university’s workings, as it’s normal for sponsors to supply it a detailed business plan.
The master’s in interdisciplinary data analysis would be a joint project of Nechyba’s institute and the Information Initiative at Duke, a group headed by computer science, engineering, math and physics professor Robert Calderbank that orchestrates “big-data” projects at the university.
Nechyba and Calderbank reckon the new degree, after ramping up, would serve 30 students in each year’s incoming class.
Cost-wise, it would cost an estimated $982,418 to launch, and by its fifth year about $3 million a year to operate. The revenue would come entirely from tuition, and the program’s sponsors are offering to share some of it with other academic departments that allow the use of their courses as electives.
That’s helped secure backing from department chairs, deans, vice provosts and other administrators, who say the extra money could alleviate bottlenecks plaguing some course offerings.
For some courses in statistics, math and programming, “the proposed funding model provides resources so the department would be able to open more sections” and handle the demand for classroom seats already being pumped up by Ph.D. and master’s students, Merlise Clyde, chairwoman of the Department of Statistical Science, said in her endorsement of the new degree.
Backers acknowledge there’s no short of data-analysis degrees at Duke. Six are already on the books, and the Fuqua School of Business is in the process of launching one of its own.
The common threat in the other offerings is that they’re either highly technical or highly specialized, tailored for instance to the needs of future Ph.Ds, economists and health-care and business executives, they say.
It should promote “broad collaboration to connect data science with real-world problems as varied as using social media to track the spread of disease, visualizing thematic trends in literature centuries or harnessing the power of electronic medical records to improve population health,” the business plan says.
At least to start, the program’s overseers will look to admit recent undergraduates and returning professionals who’ve already had coursework in statistics, basic calculus, linear or matrix algebra and computer programming.
It’s possible that later on, they’d take on students who might need some catch-up work in those areas, via so-called “boot camp” courses taken the summer before joining the program.
Participants in theory could finish in about a year, but more typically they’d need three semesters and in many cases a fourth to complete their work.
The core of the program would be a five-course sequence covering theory, a deep dive into analysis methods and the math behind them, and the tools available these days to communicate results to broader audiences.
Organizers stress that students will have to learn to work and communicate with people from a lot of academic and professional disciplines.
Not all students would necessarily pay full freight. A fair bit of the revenue that arrives via tuition would go back out again via discounts, on average 20 percent for the first three semesters and 75 percent if students stay for the fourth.
Reaction on the Academic Council so far has appeared favorable, but the group won’t vote on the proposal until next month.