The secret to being a successful data scientist (hint: it isn’t data or science)

old key in door

I wasn’t fortunate enough to be at O’Reilly’s Strata Data Conference in London earlier this year, but in his roundup of the event’s big takeaways, CIO contributor Martin De Saulles listed eight “factors shaping the future of big data, machine learning, and AI.”

No. 2 on his list is “changing skillsets for data scientists.”

“Cassie Kozyrkov, Google Cloud’s chief decision scientist, pointed out in her talk that as the UX for ML tools is improved, the skills required will become less technical and more focused on the ability of data scientists to work across silos and be more integrated into the business,” De Saulles writes.

Work across silos and be more integrated into the business. What could that mean, exactly? I had an idea, but wanted to be sure. So I Googled Cassie’s name along with “data scientist” and quickly found an interview in which she’s asked to offer three pieces of advice to aspiring data scientists. They are:

  • Useful is worth more than complicated.
  • Data quality is worth more than method quality.
  • Communication skills are worth more than yet another programming language.

That third one was what I was looking for. No. 1 also makes sense though — let’s take that one first.

“Useful is worth more than complicated” means the data has to serve some strategic purpose, whether that’s helping better understand customers, uncovering market opportunities, or identifying areas of inefficiency. Data for data’s sake is no better than technology for technology’s sake. So data scientists must understand enough about the business — often across departments — to be able to generate data relevant to the task at hand.

None of that will matter, however, if the data scientist is incapable of adequately communicating what the data means within a business context. And that brings us to “Communication skills are worth more than yet another programming language.” It takes more than a dull PowerPoint presentation crammed with thick charts to do the job. A great data scientist uses the data to tell a story, create a narrative, provide context, and create a map of possibilities (including potential problems and pitfalls).

Perhaps ironically, a successful data scientist needs to be something of an extrovert, or at least someone who can put on a convincing performance. Maybe acting and storytelling classes should be part of the curriculum for data science majors. I mean, why not? It can’t be all algorithms all the time!

Technologies may come and go, but soft skills — the ability to communicate and collaborate — will always have value. Don’t neglect them or underestimate what they can do for your career.

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.