I have started working on some open source actuarial tooling. Mostly relevant for non-life pricing.

The open tooling is quite lacking in the sense that one has to create a lot of stuff manually. Especially for the evaluations, even though most of them are quite common in the actuarial realm.

Also I wanted to tinker with Jax. So I am basing all on Jax, polars, seaborn/matplotlib and scipy.

I already have lift plots and Poisson regression in Jax. I’ll keep adding to it over the next months as time allows.

I have ported a few internal libraries to polars from pandas and had great results.

I never liked pandas much due to its annoying API, indices and single thread only implementation (we usually get a 10x performance boost at least and for me that also means improvements in productivity). Also, pandas never handled NULLs properly. This should now work with the pyarrow backend in pandas, but we can’t read our parquet files generated by PySpark. With polars it mostly just works, but we use pyarrow to read/write parquet.

Overall I can recommend it, conversion from/to pandas DataFrames was never an issue as well.

We have a 4 and a 6 year old. They are allowed a maximum of 30 minutes of screen time.

The 4 year old would get less if it weren’t for the 6 year old.

Our older one can decide between certain iOS games (like monument valley and some others) or watching a video. Grizzy and The Lemmings is something they enjoy a lot at the moment, the younger one like to watch Blaze and the older one likes to watch Checker Tobi. The latter is a fellow explaining all kinds of stuff in an age appropriate way, but 6+ would be even better.

We are planning to keep the 30min for the forseeable future. But there are exceptions like long travels, sick days, etc.