Who’s Training the Next Generation of Data Scientists?

A projected shortage of qualified data scientists could leave U.S. businesses unable to tap the value of big data. To help meet that demand, the University of California at Berkeley has developed a master’s degree program to train new data scientists.

The U.S. faces a projected shortage of 140,000 to 190,000 qualified data scientists – not to mention 1.5 million data analysts – by 2018, according to a study by the McKinsey Global Institute. IT professionals with the skills to analyze big data and guide businesses’ decision-making processes will be in high demand and short supply.

The University of California at Berkeley’s new School of Information (iSchool) masters’ level program aims to help students gain the knowledge, tools and training to land high-level, highly sought-after positions with businesses looking to use big data to improve efficiency, create new revenue streams, and compete more effectively in the marketplace.

The new Master of Information and Data Science (MIDS) program is the school’s first online-only degree program and is an effort to preemptively address businesses’ need for skilled data scientists, says Dean of the School of Information AnnaLee Saxenian.

“There certainly are folks today in business and in academia who can fill these roles. The problem is there aren’t enough of them to fill the need. The effective use of data science has applications in almost every business in every organization around the world, and that’s the issue.”
–Michael Chui, principal, McKinsey Global Institute

What’s been missing in the market, Saxenian says, is mid-level, master’s degree training that can bridge the gap between workers in business who are responsible for collecting the data and the current crop of data scientists, many of whom have Ph.Ds and are working in academia.

“There certainly are folks today in business and in academia who can fill these roles, and they are very valuable,” says Michael Chui, principal, McKinsey Global Institute, the research arm of the management consulting firm McKinsey & Company.

“The problem is there aren’t enough of them to fill the need. The effective use of data science has applications in almost every business in every organization around the world, and that’s the issue,” Chui says.

“In the future, we’ll be dealing with information not just as text and physical artifacts, but [also as] video, data, audio, sensor data collected from computers, Web clickstream data — and that will all be globally networked. We are going to need a new education paradigm to address that,” Saxenian says.

Above and beyond the new types of data, graduates of the program will be educated in the larger social, economic and personal usage issues that surround data, she says. This new degree program is much more narrowly focused on teaching students to work with data sets of all sizes; to get them to understand how to ask good questions about data; to teach them how to clean it, extract it, put it together and explore it. It’s about how to use statistics and machine learning tools, across the whole spectrum of data analytics, Saxenian says.

“We’re not just educating people in programming and data mining, but in law and legal issues, in social science and behavioral studies, user interface design and user interaction, for instance,” she says.

Where the Data Scientist Jobs Are
Businesses in almost every industry are finding ways to use data to improve efficiency, create new revenue streams, better target marketing and advertising to customers, and develop new products and services, says McKinsey Global Institute’s Chui.

“We don’t see any area of a business that can’t benefit from using data to streamline processes, better meet the needs of customers and users — to become more effective and more competitive,” Chui says.

Currently, demand for data scientists is concentrated in “traditional” sectors like IT and finance, and applicants to the program tend to hold careers in these fields, says Saxenien. But, she says, the application of data science should expand greatly over the next few years, offering graduates opportunities in real estate, government, healthcare, construction, manufacturing and more.

In addition, McKinsey’s Chui says, graduates from the MIDS and similar programs will need to educate their employers and future colleagues in how to be smarter, more effective consumers of the data that’s analyzed.

“When we studied the need for these kinds of skills, we found that businesses need executives, managers and internal data analysts to be smarter consumers of the analytics that data scientists perform,” he says. “So, these folks will have additional responsibilities to educate their future workplaces on how to ask the right questions and probe how the data can best be used within the workplace,” he says.

Data in Demand
UC Berkeley iSchool’s first MIDS class will start in January 2014, and the first graduates are expected to enter the workforce sometime in 2015, Saxenian says. While the degree program takes a year to complete, most applicants to the program currently hold full-time jobs, and will be taking courses on a part-time basis. Once they graduate, students can command a starting salary of around $100,000 or more, she says.

While UC Berkeley’s program is the first of its kind, Saxenian says she expects to see similar programs cropping up at schools across the country as demand for highly trained data scientists increases.

“These educational programs are ramping up. I think we’ll start to see more of them, and it’s going to take us as a University a while to produce these graduates. The students we have currently that are working in the data space are getting snatched up by companies, and at a starting salary for our master’s program graduates of around $100,000. That’s an incredible opportunity,” she says.