Here's my Fair 50 States, a map that I made in August/September of 2016 during my pre-election anxiety. So I made this map with only the 2008 data of DRA, and then updated it in the coming months when 2012/16 PVI was released. In this regard then, it reflects a regular redistricting process, where you map with old data and then your districts are shuffled as the coalitions change. This is best seen in the south, where I drew LA/AL/SC to have one swing seat centered on the main metro, the BVAP seats, and then the rest as geography demanded. However, when the 2016 data released, the Orleans/Metairie seat lurched leftwards because Katrina artificially lowered the cities AA totals. Since I made the map, I have updated states as I draw new maps that reflect my previous criteria better, but I always try and ignore PVI.
But, the above wasn't what I based the map on. I drew districts to preserve communities of interest, and so here was my hierarchy of criteria:
- VAP or Minority access seats. They get to avoid the rules minimizing cuts. Native reservations as also ignore the cut rule and need to be kept whole in nearly every circumstance.
- Minimize counties cut. Counties are defined as the simplest measure of communities of interest in 90% of circumstances, with exceptions for geography like CA's central valley or with cities that cross county lines like Columbia South Carolina and Arlington Texas. Another exception is New England towns which supersede counties.
- Try to base multi-county districts off of metro regions and commuter areas.
To my surprise when I finished the map and PVi was released in every state, the partisan balance was exactly 218 D 217 R. Of course that doesn't really matter because the median district is D+0.3 Wyoming valley based PA-08. There are also more Democratic districts within the <D+5 competitive band then Republican districts within the <R+5 band. I don't have the exact numbers but I recall the total number of said districts to be between 90-100. Also, one can really see the geographic biases of our nation written large on this map. Dems probably need a D+4 year to win a majority in MI, a D+2 in PA, D+6 in WI, and D+3 in MN, reflecting their geographic disadvantages in the Midwest. In Contrast, A D+3 year probably gives the Dems 17 seats in Texas to the GOPs 19, and a D+4 year of course leaves the GOP with only 8 seats in CA. Dems do better in the West/Southwest/New England then their PVI would normally allow them, and the GOP does better in the Midwest/Inland South/Florida. The Black Belt keeps the Deep south mostly in line with what one would expect from PVI, and the Mid-Altantic's Urban/Rural divides keep the states district counts within the expected range.
Overall the things that I found the most interesting during this project were: The reorientation of WI-06/07/08, The VA BVAP seats shifting westwards which allows a Chesapeake/VA Beach/Norfolk bits seat to form in the Tidewater, the LA BVAP seat without New Orleans (one can draw two BVAP seats...), and the South Brooklyn seats - keeping the Hasidim in one seat made the Staten seat barely D+ wheres the Hasidim seat is only barely R+ since you run out of compact voters and have to start grabbing from the dem-voting groups.
Also USAtlas colors because that's what I made this for.
