A 2011 McKinsey report estimated that the United States faces a shortage of 140,000 to 190,000 people with deep data analytics skills, and 1.5 million managers and analysts to make business decisions based on their findings.
In response to this shortage, GE has been training data specialists internally for the past few years, said Marco Annunziata, GE’s chief economist. In 2011, the company opened a software center in San Ramon, Calif., where hundreds of new employees were hired and trained to consult on Internet projects across the company. For instance, a specialist from the center might help employees in GE’s aviation subsidiary collect and analyze data from jet engines to improve production and fuel efficiency.
Until the global IT workforce produces enough people who specialize in both data science and software or hardware engineering, “we need to start developing them, to some extent,” Annunziata said. GE hopes to train about 1,000 of these specialists.
The company also looks for these interdisciplinary skills in new hires, Annunziata said.
“We will have more and more need for people who are a combination of data scientists and operation managers — people who have both an understanding of how to use data, how to use analytics, and also an understanding of their own business lines,” he said.
Cisco, which recently announced plans to develop “fog computing,” or a network to collect data from devices making up the Internet of Things, is also looking for similar hires, said Joseph Bradley, managing director of Cisco’s Internet of Things division. But the company is also looking for candidates who can collaborate with people in other industries, even outside the company, he said, to ensure Cisco’s networks are supported.
“If you looked 10 years ago, across enterprises, 80 to 90 percent of innovation came from within the company. If you think about that now, it’s close to 50-50. In some cases the majority of innovation comes from outside the company” as start-ups, hardware manufacturers and developers all seek to take advantage of the Internet of Things.
Each point in the network is producing large volumes of data that need to be processed in real time, and many IT training programs do not yet train graduates to analyze these streams of information, McKinsey Global Institute analyst Michael Chui said.
A handful of universities have designed data-science programs to prepare students to work on Internet of Things projects. For instance, in September, the University of California at Berkeley’s School of Information unveiled a master’s degree in information and data science. All classes are held online; the program’s first cohort is learning advanced statistics, software programming and how to process the data collected from sensors and mobile devices, among other skills. The students are also required to study ethics and data privacy.
Carnegie Mellon University, the Massachusetts Institute of Technology and Columbia University, among other schools, have introduced similar data-science programs.
Whenever large companies like Intel and Cisco mention new Internet of Things initiatives, it is a reminder that technology curriculum must evolve to meet the growing demand for IT skills, said AnnaLee Saxenian, dean of Berkeley’s School of Information.
“People need to be able to work with data — often unstructured data at very large scales, and need to be able to explore it,” Saxenian said. “Then they need to be able to communicate it with decision-makers.”
Almost all in the 28-person pilot class are working professionals who are completing the degree in their free time, Saxenian said, and it could take 12 to 18 months to do so. Most are 10 to 15 years out of college, and many have PhDs, she said.
But specialized data-science programs like these are not yet commonplace. Until IT training programs can churn out candidates skilled in both data science and engineering, those who specialize in each separately will need to work together, said Gartner analyst Hung LeHong.
“What we are recommending is for [operational technology] and IT to work together,” he said. “An IT person has excellent discipline and methodology, and can share that with their [operational technology] counterparts. IT needs to understand machines. We feel the first step is convergence, collaboration.”