UC Berkeley adds data science to its virtual classroom

Most companies are on the hunt for a data scientist, a profession the Harvard Business Review once called “the sexiest job of the 21st century.”

But university students are not graduating with the requisite skills to meet the demand.

The University of California, Berkeley, is experimenting with new courses and programs. Starting in January, Berkeley’s School of Information, the “I School,” will offer the country’s first fully online master’s degree in information and data science.

Students enrolled in the program will learn how to parse through millions of rows of data and will learn coveted skills, like data visualization, machine learning, database analysis, and data security.

And what’s another benefit to the program? Companies will know exactly what skills these Berkeley graduates possess.

Anthony Goldbloom, CEO of Kaggle, told me recently that it’s difficult for a company to recognize if a data scientist is any good — or even if candidates are data scientists at all. To address this problem, Goldbloom’s startup created its own credential and ranking system, which is already proving useful for employers.

“Companies are getting bitten by hiring a data scientist who isn’t really a data scientist,” said Goldbloom.

Why online only?

Berkeley professor AnnaLee Saxenian, who is also the dean of the School of Information, hit on the idea of a new data science master’s program several years ago. She immediately hit a roadblock. The university simply didn’t have the funding, resources, or the space for such an ambitious proposal.

Back in 2011, Berkeley and a number of other ‘public Ivies’ were in a state of fiscal peril. The collapse in state funding prompted Berkeley to double its tuition and admit students from out of state, who are willing to pay a premium.

In this climate, Saxenian had two options: She could go to alumni to raise money for a new building or offer the program online.

“We were bursting at the seams,” Saxenian recalled. Berkeley is still battling to retain its storied research and academic programs in the face of budget cuts.

She opted for online and began researching various options. She eventually found a partner in 2U, a company that works with universities to craft online degree programs.

Students will graduate with a Berkeley degree, but 2U is laying much of the groundwork. Online education is taking off at colleges and universities around the country. The two organizations agree to split the revenue and the workload.

“We do the grading and the grant the degrees; they have the tools to develop the curriculum,” said Saxenian.

More specifically, 2U is responsible for vetting the first round of applicants, designing and conveying interactive content, and providing the learning management system (LMS). 2U will also be responsible for marketing the new online degree.

Berkeley and other California universities are still experimenting with models for online learning. San Jose State University recently shuttered the five online courses it introduced in April in a high-profile partnership with ed-tech startup Udacity.

Saxenian is aware of this recent failure, but she stressed that 2U’s team are experts at engaging students with digital content. “It’s night and day” from most of the massive open online education providers (MOOCs), Saxenian said.

According to Saxenian, the MOOCs, which include Coursera and Udacity, typically offer interactive content, but students receive little or no feedback. But Berkeley has developed a “two-way delivery mechanism,” meaning that student coursework receives grades and peer review.

In addition, the Berkeley program encourages its students to meet up in their spare time.

Why 2U?

Resource-strapped universities across the country are relying on 2U and its slew of competitors. It’s a means for them to offer innovative new courses without having to devote resources to building new technology and infrastructure.

In addition, Berkeley will pocket about 40 percent of the revenue and potentially more as the program increases.

“Producing this content is expensive — it’s simply not something we had the resources to do up front,” said Saxenian.

Berkeley selected 2U because the company has worked with hundreds of other universities, and student completion rates in its programs are astoundingly high. In the past, about 80 percent of students have completed an online program.

Online education partnerships may seem like a no-brainer for universities, but it poses some ethical quandaries. Is this outsourced education? Berkeley professors are providing the course content, so students can expect to benefit from their expertise.

Saxenian stressed that the school would have complete control over the academic content — or the faculty committee would not have agreed to the partnership. This is clearly articulated in the terms of the contract.

“A vendor developing content,” she said. “Now that would worry me tremendously.”

With all the hype around online education, universities may find themselves in a precarious position. The Berkeley case shows that it’s important to set clear ground rules.

What are the alternatives for budding data scientists?

Most universities offer relevant courses for would-be data scientists: statistics or machine learning, for instance. But few have developed fully fledged courses that teach students how to wrangle data or deploy a Hadoop cluster.

Still, higher learning does offer a few alternatives to the Berkeley online program. The Insight Data Science Fellowship and Zipfian Academy are an option for those who live in the Bay Area. Anyone can sign up for Coursera’s Introduction to Data Science, which is one of the most popular programs on the site.

Enstitute is another route for high school graduates who are keen to delay or avoid college altogether. Enstitute fellows are placed in a two-year apprenticeship — most end up at tech startups, and many will work in data science teams.

Indeed, professional data scientists often make the case that they picked up their skills in the field.

Data scientist Rose Broome said she’s leery of these types of specialized courses. Instead, Broome expects that companies would prefer to hire people with professional experience who have worked on relevant projects. “In general, I’m skeptical of university data programs,” she said.

Goldbloom is more optimistic about Berkeley’s program, and she expects that other universities will follow suit. “My gut reaction is that this course offers the basis of what you need to know as a data scientist,” he said. Graduates will have their pick of jobs, according to Goldbloom, and can earn starting salaries as high as $120,000 a year.

Would you consider a career in data science? If so, how do you intend to learn the requisite skills? Let us know in the comment section below.