Dog Food and Causal Inference
The dog food was delicious, well-seasoned, and delightful
I decided to eat the dog food.
I work for Udemy, a marketplace for courses on all sorts of things — including my sort of things: statistics, “data science”, etc.
I thought: what better way to understand the instructor’s experience with the product than to try to use it? I’m not a PM or anything, but I thought it might be helpful even as a lowly data scientist. Besides, I know some things… But enough things to put a course together? Well, there was only one way to find out.
I made a course with my take on how to think about causal inference from first principles, which you can try for free if you’re one of the first 1000 people to click this link and enroll: https://www.udemy.com/course/identifying-causal-effects-for-data-scientists/?couponCode=ZACHMADEACOURSE2.
Now, this project is not without its risks. For one thing, I’m not the world’s best talker. My video production experience is limited. I could sound very dumb. But I’ve never let that stop me from talking before. [This is called “managing expectations.”]
I tried to think of a topic where I could say something interesting — a topic I had a slightly different take on from the standard course. I also wanted to write an intermediate-level course where I didn’t need to take students through the basics of probability and statistics because that’s best as its own course, I think. So, I settled on looking at how to identify causal effects from “first principles.”
By “first principles”, I mean treating the problem of causal inference not as a problem of searching for the right tool in a toolbox, but as a process of combining what we know about the world and data to learn something about causal effects. Relative to the standard course, it has a lot of coverage of partial identification approaches, including an idea that originated here: https://medium.com/good-enough-statistics/monotone-trends-for-the-difference-in-difference-setup-73e3750054cc.
In any case, if you’re interested, check it out
Thank you for reading!
Udemy course link: Identifying Causal Effects for Data Scientists
Connect at: https://linkedin.com/in/zlflynn
If you want my help with any Experimentation, Analytics, etc. problem, click here.
Zach