Synthetic control methods are a core technique for data scientists specializing in causal inference methods. One of the most popular packages to estimate synthetic controls is the CausalImpact package by Google.
In this real world data science causal inference problem, we look to see how well the Causal Impact synthetic controls could distinguish effects from noise.
Full video on YouTube: https://youtu.be/bVpctdxM3Cc
The business problem is: if we initiated brand marketing in a Latin American country, can we detect incremental traffic from the marketing campaign using synthetic controls? I demonstrate that this package gives far far too many false positives of statistical significance.
If you appreciate this content, the best way to support is some combination of following here, subscribing on youtube, or otherwise engaging with these posts and videos to let the algorithms know these are useful. 🙏
Senior data scientist and PhD economist specializing in experimentation and observational causal inference. Previously Uber, Airbnb, Stripe, the Federal Reserve, and the SEC. More about my background.
New write-ups when I finish them
I simulate a method, run it against known ground truth, and publish what happened. No schedule, and no filler between posts.
Unsubscribe anytime.