We’ve had GIS for decades. This is an easy algorithmic solve.
The simplest is the shortest-straight-line method. Draw district boundaries with the shortest straight line that divides the population appropriately.
Funnily enough, one of the biggest hurdles to algorithmic districting is the Voting Rights Act, which actually requires some level of gerrymandering to ensure representation of minorities. A algorithm may randomly split a community of color into 4 districts in violation of the VRA.
Ie, splitting a city(with a rural area in a crescent shape around it) into two equal districts down the middle each with a sizable urban and rural population(say this gave 45% rural, 55% urban in each of these districts which is pretty reasonable), vs giving the city its own district and the rural area its own district. The first option may be more “compact” but in my opinion would lead to unfair under representation of the rural voters- same as if the demographics were swapped. Districts are supposed to “represent a community” not just be compact.
We’ve had GIS for decades. This is an easy algorithmic solve.
The simplest is the shortest-straight-line method. Draw district boundaries with the shortest straight line that divides the population appropriately.
Funnily enough, one of the biggest hurdles to algorithmic districting is the Voting Rights Act, which actually requires some level of gerrymandering to ensure representation of minorities. A algorithm may randomly split a community of color into 4 districts in violation of the VRA.
Heres my example from another comment: