My first Meetup, a "Data Science" Meetup

I attended my first Meetup last week. The main event was a panel discussion with current and former "Chief Data Scientists." As a Ph.D. student considering a career in industry, I thought it was really useful. Here is an idiosyncratic review:

• For now, data scientist is a catch-all descriptor. Cheers to the supply side for exploiting this fact and capturing wage premia. But as market size increases, data scientists enter management, and management learns, differentiation may prevail.
• An understanding of identification should be useful in industry. It's neither necessary nor sufficient for success. It's more like having a really sharp knife in your kit. That said, it is very costly knife.
• It would behoove me to learn about machine learning. Aversion to black boxes kept me away. But in certain market conditions the benefits of deploying should exceed the risks. Hal Varian has some useful thoughts.
• Learning Python might turn out to be one of the best decisions I've made during my Ph.D. If you are a Ph.D. student, learn Python. At worst you'll have learned an expressive general-purpose programming language. I'll write about the best-case scenario later.
• My sense is that the number managers without a Ph.D. who can maximize the value of data scientists is small. So I think data scientists who can confidently answer "so what" questions will find a way to both leverage and appropriate the value of their human capital. (My guess is that some of Average is Over applies here.)
• Always bring business cards to a Meetup. People expect you to have them. And it's nice when someone asks if they can have one of yours. As someone who has only been in academic settings, this hasn't been necessary (we all have websites).