McKinsey Global Institute sounded the data-science-shortage alarm back in 2011, sending waves of panic through...
IT management circles. Since then, experts have debated whether the shortage is reality or myth.
Amy Gershkoff, chief data officer at Zynga and adjunct professor at the University of California, Berkeley, Haas School of Business, where she teaches a data science for executives course, falls on the reality side of that debate.
During a presentation at the recent Chief Data Officer Summit in New York, Gershkoff argued that universities aren't doing enough to educate students in analytics and data science. In her experience, the programs available are few and far between, with an even smaller number catering to undergraduate students, she said.
Among the programs that do exist, the curriculum is often light on either the technical skills or the business acumen data scientists and analytics talent are expected to have.
The top 100 universities
While conducting a study, Gershkoff found that just 29% of the U.S. News and World Report's top 100 universities globally are offering degree programs in data science. Most of the programs are located in North America, with just three located in Europe and none located in Asia. "We are part of a global economy, so it's a problem that we have such a shortage of programs in Europe and Asia," she said.
Not only are programs scarce, but the majority are only offered to graduate students, which attract a smaller number of students compared to undergraduate programs. Of the 29 programs Gershkoff identified, only six were offered to undergraduates and only three of those six offered data science as a degree program, rather than as a certification program.
Business analytics didn't fare any better. Of the top 100 universities globally, Gershkoff found that only 17 offer degrees in business analytics. Only two of those 17 universities were located outside of North America -- one in Europe and one in Asia. And none of the business analytics programs were offered to undergraduate students.
Lack of analytics talent is a global problem
According to Amy Gershkoff, only 8% of university graduates in the United States are receiving training in fields related to data science and analytics. Globally, the statistics aren't much better. In Japan, 3% of university students are receiving analytics training. In India and China, the number dips to just 1% of students.
Exceptions to the rule are Poland and Romania -- countries that have doubled down on analytics education, Gershkoff said.
The problem with curriculum
Data science or business analytics programs at top universities often provided either the technical skills or the business skills, but not both. After taking a closer look at the curricula, Gershkoff found that many data science programs appear to be computer science or engineering programs in disguise.
"A lot of the courses had names like 'software design,' 'parallel computing,' 'software development,' 'topics in computer graphics,'" she said. Courses like these develop the technical skills data scientists need, she said, but that's only half the skill set. "Most of these programs had no courses required in anything related to business strategy, communications or any of the other skills you would need to be successful as a data scientist," she said.
Business analytics programs, often found in business schools, had the opposite problem. Courses focused on business strategy, communication and critical thinking around business disciplines -- such as marketing or product management -- but they were light on technical courses.
"Many of the courses had names like, 'introduction to data analysis' and 'introduction to data visualization,'" she said. "I'm not saying those aren't great and important courses, but for many of these programs, those were the most technical courses on the curriculum."
Even Tom Davenport, the president's distinguished professor of information technology and management at Babson College, while recently arguing that the growth of data science programs is easing the analytics talent crunch, said that business schools are failing to produce "analytical managers." One reason is that many business schools have stopped requiring students to take statistics courses.
"Students weren't enjoying them that much. And, frankly, in many cases, they were not great courses in statistics and analytics. They weren't tied to actual business practices, so schools took them away," Davenport said. The result is that business students can earn a degree without substantive exposure to statistics or analytics.
That even goes for institutions as lauded as the Harvard Business School, where Davenport was a visiting professor from July 2012 to July 2013. "I think it's technically possible to get an MBA at Harvard Business School without hearing the term 'regression,'" he said.
What can be done?
Good models do exist, Gershkoff said. She pointed to Northwestern, New York University and UC Berkeley as examples of institutions that are blending technical content with real-world business applications. And Davenport said his employer, Babson College, is another institution bridging the technical and business skills.
Gershkoff also recommended four proactive steps businesses and universities can take to shore up the data science and analytics talent problem.
- Increase the number of data science and analytics programs.
- Expand enrollment to include undergraduate as well as graduate students.
- Introduce data science at a younger age, such as middle school or high school.
- Increase access to education to prepare the workforce.
More on data science and analytics talent
Best practices on building a data science team
How to attract the best analytics talent to your organization
Dropbox CEO on recruiting analytics talent