The purpose of the degree is to give students a broad exposure to working with data, explained Berkeley professor and School of Information Dean AnnaLee Saxenian, who is spearheading the program.
“The degree that we’re creating is to expose people to the entire cycle of data science from thinking about the research and why you collect data and organize it, when do you use statistics, machine learning, algorithms to find patterns, etc., to how do you communicate with business decision makers — the whole cycle,” Saxenian said.
Full-time students can complete the online program in one year, but it is largely tailored for part-time students who are currently working, according to Saxenian. “The first thing that struck me is that people are enrolling with five or more years of experience,” she said. “They’re not just engineers, either. We have people coming with a Ph.D. in educational psychology or finance.”
Approximately 30 students are enrolled for the first class and enrollment rates are expected to increase for subsequent groups. In designing the course, professors sought input from alumni who are data scientists at companies like Facebook and Yahoo to understand what businesses want. Because data science is “such a fast-changing field,” Saxenian noted, “we want to stay as close to the industry as possible.”
Even though more universities are churning out graduates with data-science degrees, an academic background is not enough, maintained Lauren Moores, VP of analytics at ad-targeting firm Media6Degrees. “Transitioning from school to a working environment can be difficult,” Moores warned. “You have to find people who will be okay in a fast-paced world, especially in ad tech, where you can’t always build the ultimate model or you’re working with a tight revenue goal.”
When reviewing resumes, the first thing Moores said she looks for is experience. “It’s okay if they don’t have experience, since we can train someone with the right aptitude, but ideally, I’m looking for someone who’s been in the real world.”
One of the skills that come from experience is being able to “tell a story,” explained Matt Curcio, chief data scientist at data-management platform provider Aggregate Knowledge. “The storytelling component is the hardest part to learn,” Curcio noted. “How do you make the data useful as opposed to just looking at the mechanics of working with data? That’s where experience comes in.”
Curcio pointed to his work on attribution models as an example. In order to quantify the value of a click and other signs of engagement, Curcio is examining numerous approaches. Given that “there are a lot of attribution vendors that talk about a one-size-fits-all model, we need to look at the data from every angle and explain why this inventory provider is undervalued, instead of just saying this ad is worth $1.11 CPM, and that’s when you weave in stories,” he said.
Having strong communication skills is key, agreed Leslie Petry, marketing director at Aggregate Knowledge. Petry said she works with the data-science department on an “almost daily” basis to create messages and launch campaigns. “As the data scientists work with people like me in marketing, we’re able to bridge that gap between just giving data versus trying to figure out what the data actually informs,” Petry said. “From a marketing perspective, while I’m getting versed in data, they’re getting versed in telling stores. We’re able to work more in concert instead of just focusing on the different things we do.”
Ken Rona, VP of audience data and analysis at Turner Broadcasting, said he looks “for someone with a good numbers sense.” In addition to having the appropriate training, data scientists need “an intuitive feel for how systems work and what things predict other things,” Rona said. “The best are people who can take numerical concepts and map them into meaning in their heads and communicate that effectively.”
In addition, as the demand for data scientists grows, companies continue to move the goal posts, commented Daryl McNutt, a former BrightRoll marketing executive who recently joined the cross-device mobile targeting firm Drawbridge. “Instead of just working with numbers, data scientists are becoming a mix of marketers plus academics plus statisticians. On top of that, the field is moving so quickly that companies are having difficulties hiring enough people who can do all that.”
Drawbridge CEO and founder Kamakshi Sivaramakrishnan, who is also a former AdMob and Google data scientist, agreed. “We hired Daryl because as we grow our need for the right messaging and other marketing services grows, but we need someone who is well-versed in data analysis and understands the research and what we do, and that is hard to find.”