Here’s a tip that might take a little pressure off the data scientist talent search: A data scientist doesn’t necessarily need to be a math wizard with a PhD or other hard science background.
In fact, that type of person might actually prove disappointing if your goal is Big Data analytics for humans, according to data scientist Michael Li.
That may seem odd, given that Li’s work focuses on exactly the kind of credentials normally associated with the term “data scientist.” Li founded and runs The Data Incubator, a six-week bootcamp to prepare science and engineering PhDs for work as data scientists and quantitative analysts.
“Putting a team of MIT-trained physicists in a role where they are constrained to use ‘simple’ models that management can understand will not be the best use of their talents, especially if they’re thirsting for a ‘deep’ machine-learning challenge,” Li writes in a recent Harvard Business Review blog post.
Li says one simple question can determine which type of data scientist you really need: “Is your data scientist producing analytics for machines or humans?”
Hard-science data scientists need to develop highly complex models. Since machines will consume their models, they don’t have to worry about the messy task of interpreting the data for humans.
If you want analytics for managers, however, you’re better off seeking someone with a social or medical science background, because their training focuses more on issues such as how and why. These are the data scientists who focus on drawing a story from the data, even if they must use simpler models. Actual humans need to understand the results, and you know what stinkers they can be.
It’s not what you think: This is not about IT alignment or something vague like that. Author Jeff Kelly, a principal research contributor at The Wikibon Project, argues that CDOs have very specific responsibilities that require finesse, communications and other soft people skills.
Specifically, he lists:
“Arguably, the CDO’s governance and people-related responsibilities are more challenging and, ultimately, more consequential than her data and technology-related responsibilities,” Kelly writes. “While the CDO role is a new one, feedback from current CDOs backs up this notion.”
Of course, it’s a false dichotomy to say that hard-math people can’t be “people persons.” I think the real issue is that hard-math PhDs are so expensive and rare that you don’t want them tackling tasks that could be handled by someone with similar, but perhaps not as specialized, skills.
If you’re interested in learning more, Matt Asay wrote a nice piece for TechRepublic that includes related statistics.
Loraine Lawson is a veteran technology reporter and blogger. She currently writes the Integration blog for IT Business Edge, which covers all aspects of integration technology, including data governance and best practices. She has also covered IT/Business Alignment and IT Security for IT Business Edge. Before becoming a freelance writer, Lawson worked at TechRepublic as a site editor and writer, covering mobile, IT management, IT security and other technology trends. Previously, she was a webmaster at the Kentucky Transportation Cabinet and a newspaper journalist. Follow Lawson at Google+ and on Twitter.