Data literacy in high demand; academia responds

Data literacy is at the heart of new computing programs at two of the country's premier private research universities. The Data Mill reports.

A new degree program at Carnegie Mellon University and an online data science training course at MIT are focused on arming those in the workforce with new skills.

On the face of it, the two programs appear to have little in common. CMU's master's degree in product management is a year-long program ideally for software engineers. MIT's Data Science: Data to Insights is a six-week course for those interested in stretching their statistical or data science muscles.

But both programs underscore the need for a data literate workforce -- within IT and outside of it. This installment of The Data Mill looks at how two premier academic institutions continue to tweak their programs to meet the data literacy demands of the modern enterprise.

"One of the things I've seen with my time in industry and my time in academia is that it takes a while for academia to respond to changes in the labor market and in the job market," said Robert Monroe, associate teaching professor of business technologies and director of the online hybrid MBA at CMU's Tepper School of Business. "But we do change and adjust."

Data science and modern applications

A six-week online course from the MIT Institute for Data, Systems, and Society (IDSS) is an advanced training ground for those looking to expand their data literacy. As part of a host of topics, the course will cover recommendation systems and network algorithms, which Devavrat Shah, co-director of the online course and one of 10 instructors, described as the crux of modern applications.

"Recommendations show up in many forms," said Shah, director of MIT's Statistics and Data Science Center and a faculty member at the MIT IDSS. "It's not just about Amazon and products or Netflix and movies."

The topic of recommendation systems is an example of a modern app that mixes big data analytics with machine learning algorithms. Shah believes recommendation engines should be built using probabilistic graphical models, which map out relationships between people and, in the case of Amazon, for example, products. "What one tries to do is one tries to learn the relationship or structure in terms of people's preferences through these probabilistic graphical models," he said.

Network algorithms use the graphical models to make inferences which, in time, lead to recommendation updates, he said. He called recommendation systems "a new class of statistical problems" -- a bridge beyond regression and hypothesis testing.

"It's a very rich problem," Shah said.

Seeking software engineers

A new master's degree at Carnegie Mellon University is aimed at meeting a surge in demand for product managers.

The role of the product manager, who shepherd products to market, has been a fixture in technology companies for decades. But that's changing, according to CMU's Monroe.

"Companies that have not historically thought of themselves as technology companies are finding they need to take more of the product management mindset and approach to building their products," he said. "As a result, we're seeing greater demand for software engineers, product managers, data scientists -- and we're seeing that demand in a much broader set of companies."

Finding candidates with a mix of the business and technology skills needed to excel in the role is difficult at best. Product managers should not only understand the business use case and act as a voice of the customer, but they also need to "work closely with engineers and product developers to bring the product to life," Monroe said. They need to be technologically competent because they guide how the product is built, advise developers on what new features to add and help set a realistic schedule for development.

"We've been calling them gems," he said. "Rare gems."

The Master of Science in Product Management, a joint program between the computer science and business schools, is an attempt to make those rare gems easier to find. Unlike the school's two-year product management MBA track, the new master's degree is a one-year program for the data literate.

"It's much easier to take somebody who's spent four years learning to be a really good software engineer ... and give them the basic business skills they need in a year or so than it is to take somebody with no technical background but good business skills and give them sufficient technical depth," he said.

Next Steps

Soft skills is now an IT job requirement

AI skills are needed in the enterprise

Building digital skills for the future

Dig Deeper on Enterprise business intelligence software and big data