TIL: Go is fast, Python is slow.

I had no idea how true this was until I built a Bayesian AB testing stats engine in Go, and compared it to python PyMC3

The repo is a python wrapper around our Bayesian-Go stats engine. You can use the repo to compare PyMC3 bayesian analysis to our API (we're about 1500x faster).

You can also use the repo as a python entry-point to our API, if you want to build your own AB experimentation tool but use our backend.

I'll spend some time on a proper SDK if there is enough interest :)