Why You Might Need a Data Science Therapist

When I became widowed 14 years ago, the first thing I did was get help. I felt unprepared to raise small kids on my own, so I was terrified and disoriented. If you’re leading a data science team for the first time, maybe you feel the same way--maybe you need a therapist too.

No, not THAT kind of therapist.

Google Chief Decision Scientist Cassie Kozyrkov ranks social scientist #8 on her list of positions to hire in a data science team. Why? Cassie’s view is that while some leaders are excellent motivators, influencers, and politicians, they’re unskilled in data-assisted decision-making.

Her suggestion: get some help from social scientists. A data science therapist, if you will. A Chief Geek of Staff.

Data science therapists have social science and analytical skills . They might have backgrounds in  behavioral economicsneuroeconomics, or “judgment and decision making” (JDM) psychology. They might be a data scientist with plus listening skills.

Like any good therapist, data science therapists guide; they don’t decide. They’re a trusted advisor, a brainstorming companion, a sounding board. They ask more questions than they answer. They distill, dig and reflect. Their role is to help clarify ideas and turn ambiguous intuitions into clear instructions. 

Data science therapists think outside the data box. Instead of working with what you have, they’ll help question the quality, integrity, and relevance of existing data. They hunt for third-party data that puts internal findings in context. And when the data is missing, they lead efforts to design surveys and experiments to explore mysteries.

I met a data therapist this week. He was a chief analytics officer. He’s binned all pre-COVID consumer preference data. Decades of consumer preference research is no longer relevant. He’s starting over. 

That’s perspective. 

Data science therapists are the straw that stirs the data science drink. They add perspective to the last mile of decision making, where the data rubber meets the business road. If you don’t have one, consider getting one.


Previous
Previous

Uber Eats for Data Science (ModelOps)

Next
Next

Reverse Mentoring