TidyTuesday python: analyzing GDPR fines
I use pandas and siuba to analyze data I've never seen before--GDPR fines over the last two years. Some key things I cover are rounding down to the nearest week (10:30), using a log scale in plotnine (20:20), and splitting + unnesting strings (40:05).
Timeline:
2:35 - create the jupyter notebook
5:52 - first plot, number of cases per country
6:25 - re-ordering pandas categorical with fct_reorder
12:00 - different ways of converting to a datetime
15:15 - why are these dates set to 1970?
17:15 - second plot, number of cases over time
23:00 - investigating GDPR fine sizes
32:00 - formatting plotnine graphs w/ log scales, angled axis text, formatting numbers
30:30 - using add_counts or fct_lump to keep groups with many rows
40:00 - splitting and unnesting strings in article_violated column
48:00 - violin plot of fines by articles violated
56:20 - parsing article id with regular expressions
01:01:40 - joining GDPR case data with GDPR text data
01:03:10 - why are there so many rows in join result?
01:06:50 - improving regular expression
Resources:
notebook code: https://github.com/machow/tidytuesday...
more analysis notebooks: https://github.com/machow/tidytuesday...
siuba: https://github.com/machow/siuba