Katie Bauer!

The surest sign that a company has no idea how to work with Data Science is requesting Insights™ as its primary output. You can tell just from the word--it's vague and sort of mystical, which is not exactly how you want to describe your quantitative teams

Yes, data analysis is about convincing someone (perhaps yourself) that something is or isn't true, but it takes a special kind of talent to do it consistently. It's easier to remember your analyses that changed someone's mind because it's easier to remember unusual events

That's definitely been the case for me, at least. I've done a lot of analysis over the course of my career, and I can only think of a handful that meaningfully changed the conversation around a subject

I've been mulling over what I think the alternative to Insights™ should be and while I have yet to come up with a satisfying conclusion, I do think one of the biggest flaws of the DS-As-Insights-Generators model is that it's a narrow strategy

If your job is to ensure data is a part of the decision making process at your company, you probably won't achieve it with artisanal Insights handcrafted by a data scientist. Maybe if you work for a magical company where the DS-stakeholder ratio is 1-to-1, and if you do--hey hmu

One of our unique strengths as DSes is getting data into a shape where the signal is differentiable from the noise, so maybe we should direct our energy towards creating environments where data is widely accessible and hard to use wrong