Quote:
Originally Posted by Steely Dan
Is that something you know how to do?
It would be interesting to get a ward map drawn of 50 roughly equal population rectangles, and then get it out on social media under the banner "chicago's new ward map, if the city wasn't governed by abject morons".
|
Looks like this problem has been studied:
https://www.brown.edu/news/2017-11-07/redistricting (and the real paper:
https://arxiv.org/pdf/1710.03358.pdf)
Their source is available, so should be straightforward to apply to Chicago. Now if only the baby was being less fussy tonight.
Snce population data is only available for census blocks, you would have to be composed of them. Still, you can represent each block by its centroid and demand convex regions on that dataset, even if the blocks will then make them a bit jagged. This appears to be what they did too, after reading it .The alternative would be to assume population is uniform within a block but that would cause all sorts of practical problems anyway.
One potential problem with this paper's approach (and any...) is that of course, the convexity of your region depends on the map projection, and they seem to just use an equirectangular projection, which means your "straight lines" are neither geodesics nor rhumb lines. Using Mercator at least would be a lot better, and it's easy to adapt their approach for that.
another edit:
Ok, their code is clunky but I figured out how to use it. Now I just need centroids and populations for every census block in the city lol. Anybody happen to have these? Seems like a bit of a pain to get it from the Census and Chicago's data web site only has it for 2010.