Pipeline to analyze Norwegian criminal justice data

The state-run broadcasting company here in Norway has been digging into our justice system and sentencing. The result is a report[1] and some stories[2][3] on how women and the rich get lower sentences in many cases.

As part of this they published the pipeline used to produce the results, which I found interesting as I have some statistics background but not a lot.

Part of the pipeline is using ChatGPT for certain tasks, like detecting aggravating or mitigating circumstances.

I also hadn't heard about R Markdown[4] before, which nice plain text nature made me discover Jupytext[5]. I had tried Jupyter Notebooks way back, but didn't like the non-plain text nature.

[1]: https://www.lighthousereports.com/methodology/norway-crimina...

[2]: https://www.nrk.no/norge/xl/de-rikeste-far-oftere-lavere-str...

[3]: https://www.nrk.no/ostfold/xl/kvinner-far-kortere-fengselsst...

[4]: https://rmarkdown.rstudio.com/

[5]: https://jupytext.readthedocs.io/