Chapter 6 - A milestone, but shall a little bit math would also help here?

Hi again @AlxndrMlk ,

I am really enjoying your book, and, again: Congratulations!

After reading Chapter 6, my impression is that the book engages in a very practical topic on how to use Causal Inference for many chapters, which is very good. However, I missed some technical details in Chapter 6 that I am trying to reproduce.

I took the chapter pynb, and combined with the chapter 7 pynb in order to reproduce the pipeline you used to demonstrate the intervention vs observation calculation of the effect size of X on Y across diverse DAGs representing the data generative method. Your construction of chapters 6 and 7 pynb were very didactic.

Although it was possible to reproduce the results for IV, Front Door and Back Door cases matching the E(Y|X) from the data generative process directly with intervention data, but very reassuring with the observational data with both the formula and the DoWhy,
I relied on the DoWhy Estimand´s expression.

I would feel more confident having learned to deduct the expressions, at least for the base cases. My gut feeling is that if I combine the data generative expressions and write in the form Y=f(X,Y), I could rearrange the partial derivatives to find the E(Y|X) expression, it worked for Front Door case, but when I wrote for IV the expression diverged from the DoWhy estimand, so I am missing some point(s).

Can you guide me on where to find the mathematical deductions of the expressions for such didactic simple cases? And maybe consider adding it in the next edition.

Thank you

Start with the Chapter 6 and Chapter 7 notebooks and compare the IV, Front Door, and Back Door examples with the DoWhy estimand expressions. Done means adding clear mathematical derivations for the base cases or documenting where readers can find them, subject to the next-edition scope.