IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

3 Key Ingredients to a Data Science Degree

Universities need a data science degree, and these ingredients go into making it.

No data science degree exists in the U.S. And that's a problem, because huge amounts of real-time data in all sectors need to be analyzed and acted on.

But this process currently passes through five people with specialized skill sets in different departments. And it takes weeks, if not months to do. Oftentimes, the market has changed by then, so the data's not even relevant.

Companies need one graduate who can do it all quickly: extract data, create tables for analysis, run the analysis, generate results and make decisions on those results. They're currently training these people in-house because universities aren't providing training.

However, in the long term, universities need to produce graduates who can fill these positions, said Jennifer Lewis Priestley, an associate professor of statistics and director of the Center for Statistics and Analytical Services at Kennesaw State University. She recommends three key ingredients to a data science degree and shares the reasons why no one's put them together.
 

1. Cross-discipline

The all-star data scientist needs knowledge in at least four areas: mathematics, computer science, statistics and a content discipline. If someone wants to study data science, they would have to cross at least two colleges within a university, if not three. 

"Universities just don't do a very good job with cross-discipline education," Priestley said.

And they don't do a good job for a number of reasons, one of them being that professors don't have knowledge of multiple disciplines. Computer science professors know their subject; they don't know statistics. An expert in biology knows neither computer science nor statistics.

"If you have a PhD in something, you've gone very deep but very narrow," Priestley said. "And so when we talk about these cross-functional degrees, it's difficult to find professors — academics — who can move in any kind of fluid way across all of these different disciplines."

This presents a challenge for universities to be relevant in this area. A degree program in data science requires the right people from each discipline who can work together, or some professors who know each of these subjects. And they're not easy to find. Degree programs that tend to be more responsive to market needs will be better positioned to be relevant in data science education, she said.
 

2. Graduate-level

This degree should be on the graduate level. Students need a knowledge base to start with in at least one of the four disciplines. Then they can learn the others from there.

Young students can't study all four things at once at an undergraduate level because they don't have the absorptive capacity to understand all the concepts, Priestley said. If students are interested in studying any of these topics eventually, she suggests that they get an undergraduate degree in math.

The path to a math degree forces students to think differently and problem solve in a structural, linear way. That is a foundational skill in computer programming: identifying a problem, determining how to translate that problem into an equation, and using the equation tools to solve a problem.
 

3. Research-oriented

Data science is rich for research right now. Our world is currently swimming in data because it's cheap and easy to capture. But the hard part is translating that data into meaningful information that can be acted upon.

For example, Southern Company oversees energy grids and sees real-time updates every second. Millions of users throughout its area  frequently generate energy usage numbers that are being captured. And the company has to figure out how to translate that data into meaningful information. 

"These are new problems, these are new challenges that nobody knows what the best practice is," Priestly said.

Someone who thinks across disciplines has an enormous opportunity to research and present papers that could solve problems. Both the public and private sectors are facing the issues of different kinds of data -- and more data overall. 

"If the private sector is trying to solve these problems, it seems like the public sector should be working even harder to solve the problem since they're supposed to be servicing the needs of the private sector," Priestly said.

This story was originally published by the Center for Digital Education.

Image courtesy of Shutterstock