Big Data at Fuqua: The Whole Beach, Not Just the Sandbox

September 6, 2017
Fuqua Study Group

It’s hard to learn how to work with big data without experiencing the sheer scale of information that can be involved.

The Master of Quantitative Management program at Duke University’s Fuqua School of Business offers students the kind of data access business students simply haven’t had until now. Providing that access – as well as the digital spaces for them to work in – is taking Fuqua’s technology support crew into what is largely unexplored virtual terrain for a business school.

“We’re providing the environment so that before students go out to work in corporations, they have a place to play around, mess up and learn from their mistakes,” said Randy Haskin, Fuqua’s associate dean for information technology. “We’re putting segments of big data sets out there, enough that they can get hands-on experience of real-world data.”

In traditional programs, students have worked almost exclusively with data stored in individual files, typically Excel spreadsheets, said Professor Ryan Burk, an adjunct professor who worked at IBM and teaches in the MQM program.

“But we are forcing students to extract data from relational databases – information separated into numerous tables that can be linked back together – to address the questions we want them to answer,” he said. “This is exactly how a lot of business data are stored, so teaching students to query these kinds of databases means they can hit the ground running.”

“Many students have these technical skills already,  they just don’t know how to apply them to business problems."

To dig into the data in their classes, students are using Jupyter Notebook – an online application through which they can write queries in multiple programming languages to retrieve the information they need, plus visualize and explain what they find.

“It’s very seamless for them,” said Joseph Conder, Fuqua’s manager of software development services. “It’s easy to get to and it’s web-based, so they can use it wherever they are.”

It’s also easy to restore order when students make mistakes.

“We can reset them to a safe point,” Conder said. If they mess up we’ll get them back to where it was working and they can try again.”

These digital environments are built around the kind of projects faculty plan to assign – projects that take an understanding of data and apply it to business.

“Many students have these technical skills already,” Haskin said, “they just don’t know how to apply them to business problems —and until now we’ve never been able to offer them the opportunity to try it.”

Real-world data can be almost anything and in any form. When almost everything individuals and businesses do is recorded as digital information, then that information can be analyzed for patterns and trends. This has created the potential for businesses to answer the kind of questions they never could before.

“When you ask most business school students to pull data, usually you’re asking them to pull specific data out of a pre-packaged system,” said Professor Jeremy Petranka, assistant dean for the MQM program. “If you’re only looking for cost of goods sold, that’s all well and good. But real world data doesn’t look like that.”

“Rather than giving them a sandpit to play in, we’re saying, ‘here’s the whole beach -- go find what we need.’”

Students bring that experience to bear in final term projects, working with companies to solve specific business problems using actual data, in whatever form it takes.

“It’ll require them to go clear some hurdles they’re not used to,” Haskin said, “but that’s the entire point.”

The program produces graduates who can help corporations figure out what questions they need to address, and who know where – and how—to look for answers.

“As the world gets messier, we’re going to need to know more in order to get answers; we’re giving students access to the ugly data,” Petranka said. “Rather than giving them a sandpit to play in, we’re saying, ‘here’s the whole beach -- go find what we need.’”