Alex Fischer

PyData Berlin Talk on our new `within` Fixed Effect Solver

Python

econometrics

PyFixest

Yesterday I was in Berlin for the day and presented at the local PyData Meetup on our new solver for fixed effects regression. It was of course a lot of fun! The development…

Token-Maxxing / Dopa-Maxxing

Usually I don’t watch much youtube (a social phenomenon I have somewhat missed out on), but here’s an interesting video about another social phenomenon I have somewhat…

New vignette: When to Use Anytime-Valid Inference

Python

pyfixest

econometrics

SAVI

anytime-valid inference

Thanks to help by Michael, the next PyFixest release will ship support for safe anytime-valid inference (SAVI). A new vignette walks through when and why to use it in…

Instacart is running fixed effect regressions with PyFixest!

Python

pyfixest

econometrics

Instacart

marketplace experiments

Ben Knight from Instacart has published a great write-up on how they use pyfixest “in production” for measuring the effects of marketplace experiments.

Fast Wild Cluster Bootstrapping in Python via wildboottest 🐍

Python

econometrics

wild cluster bootstrap

statistical inference

Aleksandr Michuda and I have just released version 0.1 of wildboottest to PyPi.

A 🐴 race: The Wild Cluster Bootstrap vs Satterthwaite-corrected Sandwich Estimators when the number of Clusters is small

statistical inference

wild cluster bootstrap

clustered errors

Satterthwaite correction

A couple of months ago, Gordon Burtch shared an excellent Twitter thread on the merits of wild cluster bootstrap inference when the regression error terms are clustered into…

Hosting a blogdown blog on github pages

blogdown

github pages

R

I have recently moved from building this blog via the distill package to blogdown. The main reason for this is that at the time of writing, distill does not support full RSS…

Testing an R package against a Julia package on github actions

R

Julia

CI

github actions

Something I have learned the hard way: every single line of code that I write should be accompanied by a unit test. Unit tests are not only useful for me as a developer to…

1000x faster Wild Cluster Bootstrap Inference in R with fwildclusterboot 🚀

R

econometrics

wild cluster bootstrap

statistical inference

When you suspect that the error terms in your regression model are correlated within clusters, and the number of clusters is small, trouble might be running at you. In such…