December 17, 2013

3 Traits That Make for a Good Data Scientist

You might have a math background, or one in finance, psychology, sociology... even exercise and health science. No matter, you might be destined for data science greatness.

That's one lesson learned for Jennifer Lewis Priestley, who said she began her academic life assuming that math students would be those most likely to have an analytical mindset. "But that's not right," said Priestley, who joined us earlier today for an A2 Radio program, "Analytics Talent: Getting Ready for 2014 & Beyond." In realizing this, she said, "I've seen my own paradigm challenged."

As associate professor of statistics at Kennesaw State University, Priestley she is also Director of the Center for Statistics and Analytical Services. She oversees the undergraduate curriculum in statistics, shepherding many students on their paths to data and analytics careers. And while she said she's seen "incredibly successful graduates" come through Kennesaw's Master's in Applied Statistics program, their backgrounds are as diverse as could be.

Good degree candidates have to be well-versed in mathematics, statistics, and computer science if they want to be data science professionals. And they need to be good writers, so much so that Priestley said she'd steer those who aren't away from pursuing a statistics degree. But just as important are three characteristics Kennesaw has seen shared among its successful data science graduates.

"We have many graduates who have gone on to now have titles of data scientist, of analytics officer (and they're making a lot more money than their professors are!)," she said.

These individuals are all:

1. Highly numerate. This is somebody who has a numerate perspective, in the sense that they can use mathematics to solve their business problems. "They understand this idea of thinking in a very disciplined, methodical, logical way." These strong quantitative skills can come from any discipline, Priestley said.

2. Able to program. The reality is that, "data is the raw material of data scientists," she said. "And the only way to extract the data, to get it into a form that can be analyzed, and then to translate it into something meaningful, requires, at the very least, some basic programming skills." Being able to program, she added, is closely related to being highly numerate in that thinking in a linear, logical, methodical way is a natural progression into programming. And, she noted, Computer Science 101 doesn't get you to this characteristic. These people get SQL, and Hadoop and Python, and understand SAS and a little bit of R -- the fundamental tools for analytics.

3. Creative and curious. "They just want to take a dataset and know what's going on... and they go well above and beyond what you assign in class," Priestley said. These are the "kids" who ask for a dataset to play with over the holidays or turn in 10 hypotheses when asked to present three, for example. Creativity and curiosity, she noted, are the same characteristics Talent Analytics and the International Institute for Analytics learned were so important among analytics professionals, as we wrote about a year ago. (See: Get Curious in 2013.)

So, tell us, do you have the stuff of a great data scientist? If not what are you missing?

- Beth Schultz, Editor in Chief, AllAnalytics.com

©