Geographic distribution of "suspicion"

[brezniczky]

Where are the suspect areas vs. not so suspicious?

Per country - a generic solution would be needed... will see what's possible.

Then they should be put on a map.

[brezniczky]

Okay, if not a map (save the cost, save the whales!) it can just be a histogram of the distribution over the

• 13 EP election districts in Poland (sheet index)
  or, the first few digits of the area code
• Counties in Hungary (originating sheet index/name)
• Top two digits of electoral ward ID (Austria)
  there also are explanatory rows stuffed into the raw datasheet

might work better/easier than expected

[brezniczky]

btw
https://pl.wikipedia.org/wiki/Okr%C4%99gi_wyborcze_do_Parlamentu_Europejskiego_w_Polsce

[brezniczky]

there are great options
https://ramiro.org/notebook/metal-bands-map/
https://medium.com/starschema-blog/draw-a-map-of-the-districts-of-budapest-using-the-overpass-api-of-openstreetmap-and-python-bd0417469935

though Google Geolocation API will likely win anyway

[brezniczky]

Well, the capital vs. non-capital comparison was easy (as well as motivated by coding accidents :) )
A quick glimpse is as
https://github.com/brezniczky/ep_elections_2019_hun/blob/master/HU/fingerprint_exploration.ipynb
and it suggests the majority of the odd votes belongs to certain areas of the capital.
Which would be a very very (tremendously, tremendously) reasonable first! target.