Skip to content

T369331 Calculate Gini coefficients

@andrewtavis-wmde for context: I coded this up shortly before my data access for my contract ended. You should be able to re-run notebook 5 exactly as before, without any previous steps other than those already documented. It should write gini_coefficients.csv into the output directory along with all the other stuff.

I did see it run, but did not have time to check the output for relevance or accuracy. For that reason, the results are not discussed in the published report. (Also, the report was already far along in the review process, and for the most part the Ginis seem less useful than the distribution plots that are already there and that were included in the report. However, calculating these coefficients is mentioned as a possible future task in section 12.1.)

The function to calculate the Ginis is copy-pasta'ed from another WMF repo. (See inline comment.)

In any case, I hope it's useful! :)

Merge request reports