I am using CausalPy (https://causalpy.readthedocs.io/en/latest/) to implement synthetic controls for Bayesian Geo-Lift.

Goal is to test a business initiative/feature (on website) in let's say 1 EU country, and use other correlated countries as control to generate synthetic counterfactuals.

I have different KPIs that I want to measure like conversion rate, sales, profits etc. I am getting a fair bayesian r2 of >0.8 while using 9 controls. I am able to get the total treatment effects but can't get a day-level synthetic and observed data in a datafrme

Question is:

  1. How can I get the above result in a data frame so I can make a dashboard (having daily impacts recorded after a test starts) out of it for users
  2. Currently I can only input 9 controls and library throws an error when I use more than 9, not sure if it's only me
  3. Placebo Test- I want to test my treatment effects statistically using placebo tests and get a p value to conclude if the results are significant. How can I do that here?

(FYI I tried using SparseSC as an alternative but Bayesian method (causalpy) gives me results that make more directional business sense)



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.