University of Vermont

New Data Science Degree Opens Up the Infinite Possibilities of Analytics

Preparing Students for Dynamic Careers

Update November 2017:

The New York Times article, Where the STEM Jobs Are (and Where They Aren't) references a Bureau of Labor Statistics forecast anticipating that 73 percent of STEM job growth will be in computer occupations in the decade ending in 2024. Read more

The article below was originally published in Summit magazine, Spring 2017

Take one look at the “hedonometer”—an empiric instrument that delivers a data visualization of the measure of happiness of a human population in real time, based on a random sampling of some 50 million tweets posted to Twitter each day—and the random acts of daily public life seem to take on a newer, deeper level of meaning.

It comes down to data—the facts and statistics that give meaning and context to everything from Walmart sales and Lake Champlain temperatures to Colorado’s coyote population and France’s Facebook habits. Described as “the new oil of the digital economy” data defines our lives more and more each day. But until now, few programs have allowed undergraduate students to practice the three fundamentals of this field (mathematics, statistics and computer science) simultaneously.
Enter UVM’s new data science degree, offered jointly by the departments of mathematics and computer science. The degree brings an interdisciplinary approach to a rapidly growing area, arming graduates with the skills to find lucrative careers among employers who are latching on to the infinite possibilities of analytics.

“It’s an amazing intersection of how to manipulate and manage large amounts of data,” says Dean Luis Garcia.

Launched in the fall of 2016, the undergraduate program synchronizes with the work already being done in the Graduate Certificate in Complex Systems and the master’s degree in complex systems and data science. The College of Engineering and Mathematical Sciences is also hiring several more faculty to support teaching and research in these areas, an indicator of even more growth ahead. 

“The past 20 or so years have led to many industries being flooded with tons and tons of data,” says Assistant Professor James Bagrow. “But the ability to understand and learn what the data contains has not been able to keep up with the flood. Now, we’re realizing this gap and starting to address it, training people with a joint set of skills—both the ability to handle the large scale of data on the computer and the ability to ask the right questions in the right way from that data.”

Students in this program are able to feel first-hand how acquiring, storing, manipulating and curating data can reveal promising patterns of human behavior. “This is a Renaissance, and it’s incredible,” says Professor Peter Dodds, who directs UVM’s Complex Systems Center, co-runs the Computational Story Lab (a group that, as Dodds puts it, produces “blatantly fun research results”) with fellow Professor Chris Danforth, and helped create the aforementioned hedonometer. “We’ve gone from being data poor to data rich, and now that we’ve figured out the little bits, such as atoms and DNA, going forward we have ‘big data,’ which is about citizens.”

Danforth elaborates, using the analogy of making weather predictions. “In about the early 70s, if you wanted to know what the weather would be like here in Vermont tomorrow, you’d call over to Ohio, assuming that things propagate east; there was little data on the Earth’s atmosphere,” he says. “But now with satellites, we can make a forecast six or seven days in advance that’s really quite accurate. Now it’s happening with human behavior because of the Internet, social networks and mobile phones. These devices and the technology are giving us observations never possible before. For people with skills in computer science and math and data, it’s a gold rush time.”
That makes students such as Sam Zonay ’18, savvy prospectors. Originally a math major, he says he fell in love with the data science course material and the idea of using computers to complement his math and statistical knowledge. “The most fascinating parts of the program for me,” he says, “have been when I get a small taste of what I hope to do one day — helping a company that I believe in and has purpose I can support with their data pools and objectives.”

As the study of big data has guided what pops up on our CNN stream and what goes into the latest Ben & Jerry’s ice cream, so, too, is it empowering the planet to become more interconnected through the telling of stories in different ways. “It’s very playful, very creative — the aspect of design is very important,” says Dodds. “And there’s incredible social good happening.”

Disease modeling from real-world flu and dengue data at UVM, for example, has revealed that substitute teachers and replacement nurses can accelerate the spread of epidemics, as an August 1 Nature Physics study explains (see page 9). Another study co-written by Dodds, Bagrow and Danforth in the Proceedings of the National Academy of Sciences spells hope for populations knitting together more; after they crowdsourced evaluations of 100,000 words in 10 languages, the authors found “the words of natural human language possess a universal positivity bias.” So, while news headlines may be negative, the work of UVM’s data science is producing more optimistic outcomes.

Those are some of the slightly less tangible results of their work so far; there are also the business advantages for graduate students including Andy Reagan, who became involved in the data science program while working on his master’s and Ph.D. in applied math and has now landed a position at MassMutual. “We’re applying data science to drive decision-making by using a combination of data and predictive models that help us analyze risk, work with high-dimensional data, and focus efforts for successful sales outcomes,” says Reagan. “This comes down to providing better overall customer experience for policy holders and greater business efficiency.”

As Danforth points out, the frontier of data science is limitless — and largely unknown by undergrads.

“Every company in the U.S. is hiring a data scientist, and yet high school students may not even realize that this area exists, that you can determine how happy we are through twitter, or predict depression from Instagram photos,” says Danforth. “UVM can truly show people what the future is going to be like.”