Follow Datanami:
October 29, 2015

Data Science Education Gets Stronger, But It’s Not There Yet

Earlier this fall, Kennesaw State University launched the country’s first Ph.D. program in data science. While the four-to-five-year program in Georgia is sure to turn out candidates for data science jobs, unfortunately, it won’t begin to scratch the surface of the current demand.

The lack of data science skills is arguably the number one problem facing the big data and data analytics community. The research firm Gartner says the current demand for data scientist exceeds the current supply by factor of three. It’s no wonder the starting salary for data scientists is well into six digits and competition for these so-called “unicorns” is so fierce.

Universities and corporations are responding to the data analytic skills in the marketplace by launching new educational programs aimed at boosting the number of qualified data scientists. One observer noted that, over the course of the past five to six years, the number of master’s-level data science and data analytics programs at American universities has gone from the single digits to about 90.

“You can’t walk onto a university campus anymore and swing a cat without somebody telling you that they have an analytics program,” says Jennifer Lewis Priestley, a professor of applied statistics and data science at Kennesaw State University, and the creator of KSU’s data science Ph.D program.

Priestly discussed KSU’s new data science Ph.D program during a session at the SAS Analytics 2015 conference held in Las Vegas earlier this week. While the master’s level programs are good, she says, the fact remains that you just can’t teach a prospective data scientist everything they need to know over 18 to 24 months, which is the typical length of a master’s level course.

“We have this dearth of people in this country, and arguably globally, who just don’t have the depth and breadth of analytical skills that are going to be needed in the next five, 10, 15 years, which is one reason why the university systems across the country responded by developing these master’s programs,” she says.

Not everybody agrees on the definition of data science, let alone whether it deserves to have its own Ph.D. program. When Priestly penned an article earlier this year about the unmet educational needs she was seeing in the analytics field, not everybody agreed with her.

The chief objection to the idea of a Ph.D. in data science appears to revolve around the idea that data science can’t be pigeon-holed into a certain set of skills, that it’s not just about whether somebody is good at math or can program computers, but that they must also possess knowledge about the particular business problem to succeed in data science.

Priestly agrees with that objection—to a point. “No matter what you’re doing—whether you’re in in finance or biology or chemistry–guess what? Everybody has to understand how to extract, transform, and load. Everybody has to understand how to clean [data]. Everybody has to understand how to do supervised learning and unsupervised learning.  Everybody has to identify what…the dependent variable looks like. Does the dependent variable make sense?

“That’s the same! Everybody has to develop those skills,” she says. “And then everybody needs to understand how to then apply a tool like SAS…You have to undertand how to use the tool to translate the data into meaningful information, regardless of the discipline.”

Even with 8,000 people graduating every year with master’s degrees in data science or data analytics, there was a need to train people with bigger and better skills. When KSU announced the introduction of its data science Ph.D. program earlier this year, it received about 100 inquiries from all over the world.

Priestly ended up with 35 completed applications for just five funded slots (IBM and Lexis Nexus were sponsors). At the end of the day, the Ph.D. program kicked off this fall with seven candidates, who will essentially be on sabbatical over the next four years.

“We believe that we did tap into an unmet need,” Priestly said. “We believe that it was the right time to start and formally launch a Ph.D. in this space.”

Only time will tell if this is the start of a new trend. It’s somewhat surprising that there hasn’t been a PH.D program up to this point. What is clear is that the number of data science training programs will continue to increase until the need is met.

Yesterday at its Premier Business Leadership conference, SAS announced its the SAS Academy for Data Science, which will train students in disciplines ranging from big data management and advanced analytics to machine learning and data visualization.

“SAS has been in the business of data science for nearly 40 years, adapting to and addressing customer needs. Right now, our customers need analytics talent,” said SAS CEO Jim Goodnight. “Employers trust SAS certified professionals not only to manage and analyze data, but also to understand the business implications, communicate results clearly and drive better decisions.”

The new SAS program will be comprised of two distinct levels, including SAS Certified Big Data Professional or SAS Certified Data Scientist. Each track will combine classroom instruction, a hands-on case study or team project, related certification exams, and coaching in a six-week immersive experience.

The idea is to give people the analytic skills needed to handle the data deluge, say David Dickey, PhD, and William Neal Reynolds Distinguished Professor in the North Carolina State University Department of Statistics. “The SAS Academy for Data Science offers tremendous opportunity for people to either enter this field or further hone their abilities,” they say in a press release.

Not to be outdone, Adobe is touting its upcoming annual collegiate competition, the Adobe Analytics Challenge. The finals, which are being held next Friday at Adobe’s Digital Marketing headquarters in Lehi, Utah, will be the culmination of an immersive class that challenges students to use data analytics to real world business problems presented by company like Lenovo, Condé Nast, and Comcast.

Last but not least, Dell announced this week that it’s giving away licenses to its Statsitica package to all universities through its Dell Statistica Free Academic Program. SAS has offered a similar program for some time.

Datanami