Hyper-Detailed 2020 election map by community

[ProgressiveModerate]

Basically an updated version of a project I worked on here:

https://talkelections.org/FORUM/index.php?topic=546308.0

For this project, each community will be between 20k and 30k people with exceptions only in very extreme cases. Communities are constructed by looking at a mix of partisanship, city boundaries, race, income, educational attainment, and a few other minor things. Communities are generally meant to be compact. Furthermore, I want to nest a whole number of communities into as many individual counties as possible (basically trying to "county-cluster" so the clusters are as small as possible. In the northeast I largely treat municipalities as counties.

Also unlike the last project, I'm willing to completely ignore precincts, especially when they're just cursed

The map on DRA is based on 2020 Pres, but I want to use these to analyze future election results.

https://davesredistricting.org/join/a16c92a4-ae1b-464f-9f81-4d9a1abc832c

[ProgressiveModerate]

States completed so far:

[smoltchanov]

I may be an idiot, but why do i see only Puerto Rico?

[Sol]

I may be an idiot, but why do i see only Puerto Rico?

You have to pan over to the mainland.

[Sol]

This looks super cool! I'm always a big fan of alternate ways of displaying data; counties can sometimes hide as much information as they show.

Also unlike the last project, I'm willing to completely ignore precincts, especially when they're just cursed

I wonder if this potentially has the possibility of compromising your data though? With such small units of analysis, and such hefty precincts in certain parts of the country, it has the possibility of throwing off the data.

[ProgressiveModerate]

This looks super cool! I'm always a big fan of alternate ways of displaying data; counties can sometimes hide as much information as they show.

Also unlike the last project, I'm willing to completely ignore precincts, especially when they're just cursed

I wonder if this potentially has the possibility of compromising your data though? With such small units of analysis, and such hefty precincts in certain parts of the country, it has the possibility of throwing off the data.

Ye that's def something I am considering. I'm doing most of my analysis at the census tract level (US Census prides itself on it's tract level data). This does however mean when it comes to election results, precinct splitting could lead to slightly inaccurate results, but it's still probably pretty close. I'm also trying to map the 2012 and 2024 elections in this project, and since precincts are different across these 3 cycles you're going to have to compromise 2 of the elections a bit anyways, so the fairest way seems to be just ignoring precinct lines across all 3 cycles and having equal statistical noise.

[ProgressiveModerate]

Added NH, both for 2012 and 2020 Pres!





Even though the statewide margins were simillar both cycles, you can def see internal coalitions changing with Dems gaining in the denser Southern poirtion of the state while Rs have gained in rural and exurban parts of upstate.

[ProgressiveModerate]

Added RI 2012!



Def shows how Dems have eroded basically everywhere in the state, with the exception of a few of the wealthiest/college ed communities.

[I spent the winter writing songs about getting better]

What exactly is the full criteria here? Because according to this my parents' house in Bismarck and my brother's less than a mile away are not in the same community but my parents are with some random rural area northwest of it 15 miles away.

[ProgressiveModerate]

What exactly is the full criteria here? Because according to this my parents' house in Bismarck and my brother's less than a mile away are not in the same community but my parents are with some random rural area northwest of it 15 miles away.

It's not perfect but I try my best. In many of the mountain states, a few suburbs/exurbs of these smaller cities are forced to be combined with very sparsely populated rurals. Laramie is another good example of a case where not all of Laramie could fit into one COI so a good chunk is in the rural COI that Laramie COI rests in.

The way I'm largely going about this project is county by county (or in some cases cluster by cluster), and trying to group areas of simillar politics/race/income/education/ect together.

[smoltchanov]

I may be an idiot, but why do i see only Puerto Rico?

You have to pan over to the mainland.

I did, but still see only country and state borders (as a sort of background). What do i wrong?

[It’s so Joever]

I did this last year as well! Would be interesting to see how we drew things differently (since that can change a lot)