Neutral map series

[Torie]

Minnesota

There are zero municipality splits (unless you count Saint Anthony, which rests in both Hennepin and Ramsey counties), three county splits (at least one of them unavoidable), and three whole-county CDs.

The overall layout was clear from the beginning. MN-05 entirely within Hennepin, box-like MN-01, big northern MN-08, et cetera. I fine-tuned it however; MN-02 previously followed the Minnesota River and Dakota County was once split in two and Washington County in three. I ultimately rotated MN-02, MN-07, and MN-06 to get what I had now. I also appreciate how  Hennepin and Wright counties together have just enough for 2 CDs between them.

There are 4 Biden seats and 4 Trump seats. 4 seats voted for the winning party by double digits - two each for both Trump and Biden. This map would probably break 4-4 most years, but Rs and Ds could win as much as 6 seats. My MN-01 took inspiration from ProgressiveModerate's one...ironically, though, I started off with it before changing it in later edits, adding it back after seeing his nationwide map.

MN-01 (Rochester): 83W; 52-45 Trump, R+7
MN-02 (Eagan): 76W; 52-46 Biden, R+1
MN-03 (Bloomington): 75W, 10B; 55-43 Biden, D+2
MN-04 (Saint Paul): 61W, 15A, 15B; 68-30 Biden, D+18
MN-05 (Minneapolis): 61W, 19B, 10H; 80-18 Biden, D+30
MN-06 (Coon Rapids): 83W; 55-43 Trump, R+9
MN-07 (St. Cloud): 86W; 65-33 Trump, R+19
MN-08 (Duluth): 86W, 53-44 Trump, R+7

DRA link

That map is a work of art and just perfect, until one's eye casts upon Sibley County, which makes the stain on Lady Macbeth's dress utterly benign in comparison. That poor small rural county is cast to the Twin Cities metro wolves, and an erose excresence to boot, that adds an erosity point under the Muon2 rules with another road cut between county seats. So to lose that crime against nature, is worth an extra muni chop, and moving a couple of counties around, in the Torieverse.

https://davesredistricting.org/join/d9f36f81-11bc-4ad1-bb94-4bacacb84f34

One can come close to avoiding an extra muni chop, but not quite there alas. In fact, to squeeze through the 0.5% rule, a precinct in Cottage Grove needed to be cut as well. The idea with chops even when necessary is to minimize the population involved. So other muni' are taken in in Washington County to minimize the size of the CG cut.

[President Punxsutawney Phil]

Minnesota

There are zero municipality splits (unless you count Saint Anthony, which rests in both Hennepin and Ramsey counties), three county splits (at least one of them unavoidable), and three whole-county CDs.

The overall layout was clear from the beginning. MN-05 entirely within Hennepin, box-like MN-01, big northern MN-08, et cetera. I fine-tuned it however; MN-02 previously followed the Minnesota River and Dakota County was once split in two and Washington County in three. I ultimately rotated MN-02, MN-07, and MN-06 to get what I had now. I also appreciate how  Hennepin and Wright counties together have just enough for 2 CDs between them.

There are 4 Biden seats and 4 Trump seats. 4 seats voted for the winning party by double digits - two each for both Trump and Biden. This map would probably break 4-4 most years, but Rs and Ds could win as much as 6 seats. My MN-01 took inspiration from ProgressiveModerate's one...ironically, though, I started off with it before changing it in later edits, adding it back after seeing his nationwide map.

MN-01 (Rochester): 83W; 52-45 Trump, R+7
MN-02 (Eagan): 76W; 52-46 Biden, R+1
MN-03 (Bloomington): 75W, 10B; 55-43 Biden, D+2
MN-04 (Saint Paul): 61W, 15A, 15B; 68-30 Biden, D+18
MN-05 (Minneapolis): 61W, 19B, 10H; 80-18 Biden, D+30
MN-06 (Coon Rapids): 83W; 55-43 Trump, R+9
MN-07 (St. Cloud): 86W; 65-33 Trump, R+19
MN-08 (Duluth): 86W, 53-44 Trump, R+7

DRA link

That map is a work of art and just perfect, until one's eye casts upon Sibley County, which makes the stain on Lady Macbeth's dress utterly benign in comparison. That poor small rural county is cast to the Twin Cities metro wolves, and an erose excresence to boot, that adds an erosity point under the Muon2 rules with another road cut between county seats. So to lose that crime against nature, is worth an extra muni chop, and moving a couple of counties around, in the Torieverse.

https://davesredistricting.org/join/d9f36f81-11bc-4ad1-bb94-4bacacb84f34

One can come close to avoiding an extra muni chop, but not quite there alas. In fact, to squeeze through the 0.5% rule, a precinct in Cottage Grove needed to be cut as well. The idea with chops even when necessary is to minimize the population involved. So other muni' are taken in in Washington County to minimize the size of the CG cut.

Comes to show how different criteria can result in slightly different maps. In my New York map, I ended up splitting Victor Township in Ontario County for population reasons (had to stay within the 1% deviation band), but that was something of a pained decision. Here, more emphasis on municipal and county integrity lead to me shelving this, allowing me to unify Washington (shifting the Ramsey CD in a westernward direction, making both more compact). Since a chop somewhere was inevitable, it made most sense to do it at a place that would make it the most compact. (Compactness was also reason to avoid any kind of tail for MN-01) My prioritization of this led to slightly higher compactness and split scores, which is to be expected I guess.

[Torie]

Interestingly, the MN-01 tail does not generate an extra erosity penalty point under the Muon2 rules. It's a freebie. What you did with Sibley does, as well as another penalty point for MN-02 straying outside the MSA counties.

[President Punxsutawney Phil]

Interestingly, the MN-01 tail does not generate an extra erosity penalty point under the Muon2 rules. It's a freebie. What you did with Sibley does, as well as another penalty point for MN-02 straying outside the MSA counties.

This makes for an interesting test case as to how the existence of these kinds of rules changes the optimal map.
I can see why Sibley might create an additional erosity point. In terms of the overall district shape, it is a significantly larger protusion west than two counties in the larger MN-01 would.

[Torie]

Interestingly, the MN-01 tail does not generate an extra erosity penalty point under the Muon2 rules. It's a freebie. What you did with Sibley does, as well as another penalty point for MN-02 straying outside the MSA counties.

This makes for an interesting test case as to how the existence of these kinds of rules changes the optimal map.
I can see why Sibley might create an additional erosity point. In terms of the overall district shape, it is a significantly larger protusion west than two counties in the larger MN-01 would.

Just so you know, you generate erosity penalty points when the highway that one takes as the most direct route between two county seats crosses a district boundary. It is a very clever algorithm actually, that makes a lot more sense than the statistical/geometric tests that computers spit out that nobody really understands, and don't make sense if you have natural barriers. You don't put Inyo County with Tulare County for a reason.

[President Punxsutawney Phil]

Interestingly, the MN-01 tail does not generate an extra erosity penalty point under the Muon2 rules. It's a freebie. What you did with Sibley does, as well as another penalty point for MN-02 straying outside the MSA counties.

This makes for an interesting test case as to how the existence of these kinds of rules changes the optimal map.
I can see why Sibley might create an additional erosity point. In terms of the overall district shape, it is a significantly larger protusion west than two counties in the larger MN-01 would.

Just so you know, you generate erosity penalty points when the highway that one takes as the most direct route between two county seats crosses a district boundary. It is a very clever algorithm actually, that makes a lot more sense than the statistical/geometric tests that computers spit out that nobody really understands, and don't make sense if you have natural barriers. You don't put Inyo County with Tulare County for a reason.

Oh, interesting.
That's clever.
For that matter, I agree I would not put Inyo and Tulare as a singular connection. On the other hand, Kern is quite possible, all other things being equal. Inyo-Mono-Kern was a nesting I used in a 250 seat CA map thing I was doing sometime ago. (Iirc, Kern by itself was not well positioned, so Mono and Inyo helped even things out and prevent Kern from having to be paired with another county).

[Libertas Vel Mors]

This is a map of all the states done so far. The districts are colored based on how I assume they would have voted in the 2022 midterms. In Florida, I assume a lot of close holds for Ds, while Rs win by huge margins in their strongholds. Rs win Biden-voting FL-14, in what would be the most marginal Biden seat in the state.

I count 138D, 141R, and 156 thus far undone.

EDIT: I was not aware of how badly Ds did in Nassau. I thought R candidates won by narrow margins and felt a CD 1/4 in NYC would be unlikely to elect a Republican. I guess you can flip both the red Nassau seats to Republicans...

Would you mind sharing a base copy of the template you used to make this, and a guide to doing so?

[President Punxsutawney Phil]

Texas

Given how many counties Texas has, I figured nestings was a no-brainer at least when doing some more urban areas. This can be seen in the Dallas-Fort Worth Metroplex, where there are four sets of county clusters. Dallas-Rockwall-Kaufman-Ellis has 4 CDs, Tarrant-Johnson-Somervell has 3, and Collin+a bunch of counties to its east and Denton+a bunch of counties to its west both have 2.

Within these clusters, I sought to make districts as compact as practicable while still respecting municipal borders. One impact from this was drawing of a district combining Kaufman, Ellis, and as many black precincts as possible, as it was deemed necessary to have the "South Dallas" seat be black-performing. TX-30 may sit at only 37% black under these lines, but that 30% Latino minority is not going to turn out nearly as much, so you may as well consider it a strong black-plurality CD. Biden+18 should be enough to make it Solid D.

The Latino CD within Dallas County is still numbered TX-33. Its western border ends where the strongly Latino precincts in the western half of Dallas end. It will be hard for it not to elect a Latino, but stranger things have happened. TX-32 is centered on Dallas itself, but also picks up Addison and Richardson. Finally, TX-05 is Latino-plurality and centered on eastern Dallas County, taking on a leftovers kind of character. It would be fairly marginal, but D-leaning.

In Tarrant County, TX-06 is still centered on Arlington, but moves west from there, not east. A slim majority of its population lives in Arlington itself. Its exurban portions render it R-leaning, but its R lean shrinks nonetheless. TX-12 is an effort at a CD centered on Fort Worth itself, but it takes in Halton Hills to reach quota while keeping both it and TX-24 more boxlike, and to boost its function as a minority access seat. TX-24 itself fills the rest of Tarrant County left over.

In the Northern part of the Metroplex, Collin and Denton form boxlike CDs that determine the shape of the other seat in the cluster. I live in TX-03 under these borders, and have done so for about a slim majority of my life at this point. TX-03 has a higher Asian % than any other CD in the area and voted for Biden. TX-26, meanwhile, voted for Trump by low single digits. The other two CDs within these nesting clusters are extremely conservative and exurban or straight-up rural in character. TX-13 in particular comes within 60 miles of New Mexico, taking in most of the least populous sections of Texas.

Rounding out West Texas is a pair of whole-county CDs: TX-11 (containing Abilene and the cities of the Permian Basin), and TX-19 (focused on Lubbock and Amarillo). If combined with TX-13 and TX-23, they would have a majority of the state's land area, despite having barely 10% of the state population.

On the border there are five CDs, just like on the current map. However, TX-16 is shaped with municipal integrity in mind and contains all of El Paso, while TX-23 was influenced by the cluster boundaries I came up with - producing lines that encouraged them to be more within Bexar than they otherwise would be. In general, incompact fajitas are basically gone whenever they can be replaced with compact ones. TX-21 takes in a lot of the Texas Hill County and then takes from Bexar to make up the rest.

I use highways to divide up San Antonio, with Loop 1604 providing a lot of TX-21's border within Bexar. Since TX-23 ought to be Latino performing, it avoids the more whiter exurban areas in Bexar's north and instead takes in a big chunk of more urban parts of San Antonio instead. TX-20 has no rural element and follows the inner-ring Loop 410. It departs from this area to take in whiter places that have nowhere else to go. All of Bexar that cannot go into any of these prior districts goes in TX-28.

Hidalgo County is split in 2; one district (TX-34) taking in all the coastal counties within the cluster grouping I worked out, another (TX-15) taking in the rest, lest TX-34's share of Hidalgo. Corpus Christi ends up in a TX-27, running along the coast. It eventually takes the excess from Fort Bend County, establishing TX-22 as a one-county CD.

Hays and Travis are perfect for 2 districts, and so there is a new CD (TX-38), technically Latino-plurality, in the eastern half of Travis, and then there is TX-10, covering the rest of Travis and all of Hays. Adjacent is TX-35, now a whole-county CD entirely outside of Bexar and Travis, and going as far north as Bryan-College Station. Compactness concerns prevented me from considering placing the southern ends of TX-35 with Corpus Christi. Also "deep in the heart of Texas" is TX-31, designed to be as compact a district as possibly by taking almost all of Killeen (I wish it could be all of it) and some rural precincts in Bell County.

The nestings ordained the creation of a district sitting in between the Austin and DFW metros, and I numbered it the 25th, after the CD already existing between Austin and Fort Worth. While it could not take all of McLennan, it could take all of the city of Waco itself. Finally, TX-17, losing Texas A&M, gains a border with Louisiana. Meanwhile, TX-01 is one of the relatively few whole-county CDs on the map - just one of four. It gains a border with Oklahoma as TX-04 retreats due to Collin County population growth shrinking TX-03.

In Southeast Texas, Galveston, Brazoria, and Chambers form a whole-county CD. I sought to draw another whole-county CD in the area but I could not find any that made sense. TX-36 largely keeps its 2013-2023 rural territory, gaining a small slice of Montgomery County. It loses most of its share of Harris County, as a new seat is added there. The primary function of the remaining share of TX-36 within Harris County is to make it easier to draw clean minority-performing seats in Harris County.

The current VRA seats within Harris County are extremely uncompact and cannot be justified under these sets of rules. I, just like in San Antonio, used highways and other important roads as guides. The new TX-37 is bounded by the Harris County line, Spencer Road, and FM 1960 for the vast bulk of its borders. This left a large area (namely, Kingwood) that was heavily white but could not be placed within a white seat because that would be horridly uncompact.

To solve this issue, I essentially merged the eastern part of TX-02 with the northern part of TX-18. It also had to pick up many Latinos as a matter of inevitably. The new TX-02 is more-or-less the successor to the old TX-18, with southern borders reaching all the way to the Houston Ship Canal in order to take some heavily black precincts. Since this cut off the old TX-29's northern portion from its southern portion and I felt it would be futile to try to save this arrangement, I instead drew the nucleus of the former, numbered it the 18th, and took it west, where it gained a lot of Hispanic voters. This is the most securely Latino CD in Harris County and it has almost three times as many Latinos as it has blacks.

The 29th numbering is given to a seat centered on SW Harris County and considerably stronger for Rs than any other minority-plurality CD in Harris County. It has 5 Latinos for every 3 whites. I would not necessary consider it Latino-performing in a general sense, but it is quite winnable for the Latino candidate of choice in good enough circumstances. Bordering TX-29 is TX-09, which, now confined to Harris County, hoops around to take 97% of the Fort Bend-Harris county line. It remains black-performing, taking almost all the black precincts left. Latinos almost outnumber blacks here, and blacks are twice as numerous as whites - a recipe for a CD still controlled by the black community.

Finally rounding out Harris County is TX-07. It is a diverse seat, with a very strong white plurality. Its Latino % is fairly high on paper, a function of it having taken some heavily Latino precincts that could not go elsewhere. This seat is probably ground zero for pro-Dem trends in Harris County and it voted for Biden by 21 points.

This map, because it makes no attempt to compensate Republicans for their bad geography, gives them a clear minority of CDs. Biden won 21 districts out of 38. Seat number 20 for the GOP would be a single-digit Biden seat. I am not sure if Abbott won a majority of these seats in November 2022, but under a statewide uniform swing, it looks likely. Still, it would be less seats than he would have gotten in a map drawn with criteria that corrects for geography. Ds probably would be able to generally have a small majority of the state's congressional delegation under these lines.

TX-01 (Tyler): 60W, 19B, 17H; 72-27 Trump, R+25
TX-02 (Houston): 49H, 31B, 17W; 64-35 Biden, D+14
TX-03 (Plano): 48W, 24A, 14H, 12B; 51-47 Biden, R+3
TX-04 (Wylie): 66W, 17H; 72-27 Trump, R+27
TX-05 (Garland): 43H, 30W, 19B; 53-46 Biden, D+1
TX-06 (Arlington): 43W, 27H, 21B; 52-47 Trump, R+6
TX-07 (Houston): 40W, 37H, 12A, 11B; 60-39 Biden, D+8
TX-08 (Conroe): 58W, 26B, 10B; 70-29 Trump, R+25
TX-09 (Houston): 35B, 35H, 18W, 11A; 69-30 Biden, D+19
TX-10 (Austin): 59W, 26H; 61-37 Biden, D+9
TX-11 (Midland): 48W, 42H; 76-23 Trump, R+29
TX-12 (Fort Worth): 39H, 34W, 22B; 58-41 Biden, D+6
TX-13 (Wichita Falls): 69W, 20H; 79-20 Trump, R+33
TX-14 (League City): 50W, 28H, 15B; 61-37 Trump, R+14
TX-15 (McAllen): 88H; 53-46 Biden, D+8
TX-16 (El Paso): 81H, 13W; 67-32 Biden, D+14
TX-17 (Lufkin): 63W, 20H, 14B; 76-23 Trump, R+28
TX-18 (Houston): 55H, 20B, 18W; 59-39 Biden, D+9
TX-19 (Lubbock): 49W, 39B; 72-27 Trump, R+26
TX-20 (San Antonio): 72H, 19W; 66-32 Biden, D+16
TX-21 (San Antonio): 53W, 35H; 61-38 Trump, R+16
TX-22 (Sugar Land): 29W, 25A, 23H, 23B; 56-42 Biden, D+4
TX-23 (San Antonio): 69H, 23W; 51-48 Biden, D+1
TX-24 (Fort Worth): 58W, 20H, 11B; 59-40 Trump, R+15
TX-25 (Waco): 60W, 23H, 12B; 70-29 Trump, R+24
TX-26 (Denton): 52W, 21H, 13B, 13A; 51-48 Trump, R+7
TX-27 (Corpus Christi): 51H, 39W; 62-37 Biden, R+14
TX-28 (Laredo): 70H, 19W, 10B; 57-42 Biden, D+8
TX-29 (Houston): 55H, 27W, 12B; 51-48 Biden, EVEN
TX-30 (Dallas): 37B, 30H, 30W; 58-40 Biden, D+7
TX-31 (Killeen): 50W, 25H, 15B, 10A; 51-47 Biden, R+3
TX-32 (Dallas): 47W, 26H, 18B; 61-37 Biden, D+9
TX-33 (Irving): 53H, 20W, 14B, 12A; 65-34 Biden, D+13
TX-34 (Brownsville): 90H; 57-42 Biden, D+12
TX-35 (College Station): 52W, 34H; 64-35 Trump, R+18
TX-36 (Beaumont): 55W, 22H, 19B; 70-29 Trump, R+22
TX-37 (Houston): 44W, 28H, 17B, 10A; 57-42 Trump, R+12
TX-38 (Austin): 42H, 37W, 13B; 79-19 Biden, D+29

DRA link

[President Punxsutawney Phil]

This is a map of all the states done so far. The districts are colored based on how I assume they would have voted in the 2022 midterms. In Florida, I assume a lot of close holds for Ds, while Rs win by huge margins in their strongholds. Rs win Biden-voting FL-14, in what would be the most marginal Biden seat in the state.

I count 138D, 141R, and 156 thus far undone.

EDIT: I was not aware of how badly Ds did in Nassau. I thought R candidates won by narrow margins and felt a CD 1/4 in NYC would be unlikely to elect a Republican. I guess you can flip both the red Nassau seats to Republicans...

Would you mind sharing a base copy of the template you used to make this, and a guide to doing so?

Same map template I have used for many years. I actually have a secondary version including Puerto Rican municipalities. I've actually forgotten where I got them from...

[President Punxsutawney Phil]

Midwest: 8/10 (31/43)
Northeast: 7/8 (49/21)
South: 13/15 (64/82)
West: 6/11 (17/17)
Nationwide: 34/44 (161/163)

Very possible that we might end up with more Biden seats than Trump seats in the wider country outside of CA. But a lot depends on the remaining maps.

[President Punxsutawney Phil]

Utah

There was an obvious whole-county CD in form of Davis-Weber-Box Elder-Cache-Rich. That seat was always going to be drawn once I discovered that it was possible. The real question was what to do with the rest of the state. I had this, but...I disliked the unnecessary county split that it had, so I looked for alternatives. I found something satisfactory. The municipality of Murray in Salt Lake County is split for compactness reasons. There are zero avoidable county splits and two whole-county CDs. All districts on the map voted for Trump, but two of them are vulnerable and won by margins much smaller than the state as a whole.

UT-01 (Ogden): 79W, 13H; 62-32 Trump, R+19
UT-02 (West Valley City): 67W, 21H; 50-44 Trump, R+5
UT-03 (Provo): 79W, 13H; 68-25 Trump, R+27
UT-04 (Salt Lake City): 76W, 13H; 50-46 Trump, R+4

DRA link

[President Punxsutawney Phil]

Washington

This is very similar to a non-partisan map I made about a week ago. Washington was going to be the first state I did and I was going to use that map, but I felt I needed to make a new one optimized for my criteria. I finally am getting around to doing WA now.

Like in the OTL commission map and my non-partisan map, all of Eastern Washington less Chelan and Kittitas counties is an area in which two seats are nested. I drew them with both seats bordering Idaho in a boxlike style, for maximum compactness. In general, there is emphasis on not splitting municipalities, but since I dislike having to use precinct splitting, precinct borders make it difficult to keep too many of them whole. These seats do keep to municipal lines when possible and keep county splits to a reasonable minimum. For example, WA-07 is now Seattle, Medina, Hunts Point, Yarrow Point, Clyde Hill, and Mercers Island.

There are seven Biden seats and three Trump seats.

WA-01 (Bellingham): 75W, 11H; 54-43 Biden, D+3
WA-02 (Everett): 60W, 20A, 11H; 69-28 Biden, D+19
WA-03 (Vancouver): 75W, 11H; 50-46 Trump, R+5
WA-04 (Yakima): 54W, 37H; 55-41 Trump, R+10
WA-05 (Spokane): 76W, 11H; 54-42 Trump, R+9
WA-06 (Olympia): 73W; 56-40 Biden, D+6
WA-07 (Seattle): 60W, 21A; 88-10 Biden, D+38
WA-08 (Auburn): 65W, 16A, 11H; 54-43 Biden, D+3
WA-09 (Bellevue): 42W, 27A, 15H, 13B; 68-29 Biden, D+18
WA-10 (Tacoma): 59W, 13H, 12B, 11A; 57-40 Biden, D+6

DRA link

[President Punxsutawney Phil]

Connecticut

These counties are due to be replaced in not too long, but since they are the county units that CT is currently using, I am obliged to use them.

I did declare that town splits were verboten here. After toying with CT-04, I found the coastal counties in Fairfield County were too populous and I found it more fruitful extending CT-04 inland so that instead Bridgeport was placed with New Haven. CT-05 ended un eating a lot in New Haven County because CT-03 withdraw from inland areas largely because of this. Meanwhie, CT-02 took the town of Rocky Hill to reach quota. After most of Hartford County ended up in CT-01, the rest was given to CT-05.

This map largely ended up fairly well for Republicans, and CT-05 would be far more marginally Dem here. Under these lines, Rs would have certainly won it in 2022, considering how close they got in OTL. Nonetheless, all districts here are Biden districts, and Ds might even be easily favored in a neutral year to sweep, if they get an entrenched incumbent. The new CT-03 is technically majority-minority in total population, though this is a mere by-product of my mapping decisions and not something I set out to do. It remains 54% White VAP.

CT-01 (Hartford): 53W, 21H, 19B; 66-32 Biden, D+15
CT-02 (Middletown): 77W, 10H; 55-43 Biden, D+3
CT-03 (Bridgeport): 50W, 23B, 23H; 65-34 Biden, D+13
CT-04 (Stamford): 66W, 18H; 62-37 Biden, D+9
CT-05 (Waterbury): 71W, 16H; 50-48 Biden, R+3

DRA link

[President Punxsutawney Phil]

Midwest: 8/10 (31/43)
Northeast: 8/8 (54/21)
South: 13/15 (64/82)
West: 8/11 (24/24)
Nationwide: 37/44 (173/171)

The Northeast is the first region to be completed. This is not unexpected. Biden has now taken a lead in the overall national seat count for the first time.

[President Punxsutawney Phil]

Virginia

This is the first one I made. What made me pass on it was the seemingly low compactness score...just about 44 or something. For comparison, I looked at the OTL court master map and I saw it scored only 46. Upon redrawing, I ended up with 47, which is not too bad.

In the redraw, I changed the Hampton Roads area the most, seeking to boost compactness by giving the Delmarva Peninsula to VA-01. This is not unpecedented, though it has not been normal in this century. In the 1980s, Accomack and Middlesex counties were in the same district, and this seems to have been normal until recently. My redrawing of the Hampton Roads allowed me to get rid of a county split, creating a whole-county CD in the Richmond area. I'm aware that there is no seat that is over 40% black, but there are in fact four seats that are over or almost 30% black and D-favoring, which should in practice serve as black-influenced seats. Thus, I am not concerned about minority representation here.

As for the Shenandoah Valley, while I admire the VA-06 on the OTL map, it's unfitting for this series because of how narrow it is. The northern parts of it go in with VA-10, while it becomes more box-like with the addition of Nelson, Amherst, and Bedford counties, as well as Lynchburg City. This whole-county CD, one of three on the map, helps improve overall compactness scores and county integrity at the same time.

VA-05 remains centered on its particular stretch of VA. It takes generally whiter parts of Chesterfield County to boost the black % in VA-03 and create a compact border between the two districts. It continues to stretch from Northern Virginia to the North Carolina border, and it continues to lean Republican, though under these borders it becomes significantly more winnable for Ds in a good year. The Chesterfield County territory not only makes VA-03 blacker, it makes the trends in VA-05 better for Ds, as this is a D-trending county. This is even more beneficial for Democrats considering rural VA's trends opposite to the state as a whole - and VA-05 has a ton of these rurals, and VA-03 is seeing black population drain from rural areas.

In Northern Virginia, there is a nesting for 4 districts that I created, and kept in both maps. The main difference was how I drew the outer suburbs - Prince William, Loudoun, and Manassas. In this second map, I have a more box-like VA-10 with Prince William and Loudoun being entirely separate. The "tail" of the county cluster ends up with the former, the best option. Since VA-10 is just under, and Fairfax+independent cities within its borders is just over two seats, it is natural for a very small share of the latter to be paired with the former. McNair's population meant it was just right for this.

There are eight Biden seats and three Trump seats. However, in 2021, Youngkin won six seats, the median seat actually voting for him by a larger margin than the state as a whole. If I had to guess, it would split 7D-4R, with Sparnbarger still running (and winning) in VA-07 and Wexton narrowly losing VA-10 because of the conservative Shenandoah Valley territory added in. Three districts are majority-minority, two of them in NoVa, though I did not consider the creation of majority-minority seats in how I drew the region.

VA-01 (Newport News): 54W, 33B; 52-46 Biden, EVEN
VA-02 (Virginia Beach): 52W, 31B; 58-40 Biden, D+6
VA-03 (Petersburg): 49W, 39B; 54-45 Biden, D+2
VA-04 (Richmond): 55W, 30B; 62-37 Biden, D+10
VA-05 (Charlottesville): 70W, 19B; 51-47 Trump, R+5
VA-06 (Roanoke): 77W, 13B; 59-39 Trump, R+13
VA-07 (Dale City): 44W, 23H, 22B, 10A; 58-40 Biden, D+6
VA-08 (Alexandria): 46W, 19A, 19H, 16B; 72-26 Biden, D+21
VA-09 (Blacksburg): 83W, 10B; 69-30 Trump, R+22
VA-10 (Leesburg): 63W, 15A, 12H; 51-47 Biden, D+2
VA-11 (Arlington): 53W, 22A; 16H; 74-24 Biden, D+23

DRA link

[President Punxsutawney Phil]

Michigan

My goals coming in were what you might expect - split no townships, keep county splits to the minimum, and create compact districts. To this end, I did nesting. Among other things, I found you could nest three seats in Wayne-Washtenaw-Monroe and St. Clair-Macomb-Oakland. Before I did this, of course, I drew the Shark (Upper Peninsula) and the upper parts of the Mitt (Lower Peninsula) into MI-01. The shape I gave it put it almost four thousand below quota, or -0.52%, not a lot.

After this, I drew Western Michigan. Manistee, Wexford, Osceola, Lake, Mason, Mecosta, Newaygo, Oceana, Muskegon, Kent, Ottawa, Barry, and Allegan into a two-CD cluster and then made out of it two compact CDs. Trump won this Grand Rapids CD by 21 votes, making it likely the smallest numerical vote margin for any CD in this project in terms of 2020 presidential vote. Unfortunately, I could not draw Berrien+Van Buren+Cass+Kalamazoo+St. Joseph+Calhoun less its five easternmost townships+Branch was too big for a CD; it would make for a CD whose border was three straight lines on land, the ultimate box-like CD. No matter; I simply detached two townships into MI-08.

The Washtenaw-Wayne-Monroe cluster forced MI-07 east and MI-08 south, to fill in the void. MI-08 has become marginally less Democratic and on these lines, Rs probably win it by a small margin in 2022. North of MI-08 are two whole-county CDs - MI-04 and MI-05 - both of which are Republican-leaning at least on paper. However, MI-05 has gained a lot of moderate Lansing suburbs and this is one of the better ways it could be shaped from a Dem POV.

MI-08 turned out to have too much people, forcing me to expand the potential St. Clair-Macomb-Oakland cluster. Tyrone Township in Livingston County was put in MI-11, in effect turning a 3-CD cluster into a 5-CD one. Within Oakland County, I went up in columns, from Novi to Holly, Lovi to Groveland, et cetera. I finally hit a snag when I reached Pontiac. I then went east, finding that Oak Park gave me all the population I needed for MI-11 to reach quota. I resumed columns again when drawing the 9th, placing all the Macomb County townships that bordered Oakland County in MI-09. When this proved to be too much, I found that removing Sterling Heights and Utica got the numbers right.

Finally, within the Washtenaw-Wayne-Monroe cluster, I sought to make MI-07 be as white as possible in order to make the other two seats be able to fulfill the goal of minority representation. I feel that was achieved, with two districts with a strong black presence exceeding 40%, and Detroit split along Livernois Avenue. Detroit is the only municipality to be split in Michigan.

There are five Biden seats and eight Trump seats, but Ds have more safe/likely seats than Rs do. The median seat voted Trump by less than 2 points, fairly good for them considering the bad geography they have got here. In 2022, this map likely would have yielded a 7D-6R result, just like OTL.

MI-01 (Marquette): 89W; 59-39 Trump, R+13
MI-02 (Muskegon): 82W; 59-40 Trump, R+13
MI-03 (Grand Rapids): 74W, 10B, 10H; 49.013-49.008 Trump; R+4
MI-04 (Saginaw): 84W; 59-39 Trump, R+12
MI-05 (Flint): 77W, 14B; 50-49 Trump, R+3
MI-06 (Kalamazoo): 76W, 12B; 51-47 Trump, R+4
MI-07 (Ann Arbor): 72W, 11B; 60-38 Biden, D+9
MI-08 (Lansing): 80W; 51-47 Trump, R+5
MI-09 (Warren): 70W, 13B, 10A; 54-45 Biden, D+1
MI-10 (Sterling Heights): 79W, 12B; 56-43 Trump, R+9
MI-11 (Farmington Hills): 69W, 17B; 55-44 Biden, D+3
MI-12 (Detroit): 47W, 46B; 72-27 Biden, D+21
MI-13 (Detroit): 43B, 42W; 70-29 Biden, D+20

DRA link

[palandio]

Virginia

This is the first one I made. What made me pass on it was the seemingly low compactness score...just about 44 or something. For comparison, I looked at the OTL court master map and I saw it scored only 46. Upon redrawing, I ended up with 47, which is not too bad.

In the redraw, I changed the Hampton Roads area the most, seeking to boost compactness by giving the Delmarva Peninsula to VA-01. This is not unpecedented, though it has not been normal in this century. In the 1980s, Accomack and Middlesex counties were in the same district, and this seems to have been normal until recently. My redrawing of the Hampton Roads allowed me to get rid of a county split, creating a whole-county CD in the Richmond area. I'm aware that there is no seat that is over 40% black, but there are in fact four seats that are over or almost 30% black and D-favoring, which should in practice serve as black-influenced seats. Thus, I am not concerned about minority representation here.

As for the Shenandoah Valley, while I admire the VA-06 on the OTL map, it's unfitting for this series because of how narrow it is. The northern parts of it go in with VA-10, while it becomes more box-like with the addition of Nelson, Amherst, and Bedford counties, as well as Lynchburg City. This whole-county CD, one of three on the map, helps improve overall compactness scores and county integrity at the same time.

VA-05 remains centered on its particular stretch of VA. It takes generally whiter parts of Chesterfield County to boost the black % in VA-03 and create a compact border between the two districts. It continues to stretch from Northern Virginia to the North Carolina border, and it continues to lean Republican, though under these borders it becomes significantly more winnable for Ds in a good year. The Chesterfield County territory not only makes VA-03 blacker, it makes the trends in VA-05 better for Ds, as this is a D-trending county. This is even more beneficial for Democrats considering rural VA's trends opposite to the state as a whole - and VA-05 has a ton of these rurals, and VA-03 is seeing black population drain from rural areas.

In Northern Virginia, there is a nesting for 4 districts that I created, and kept in both maps. The main difference was how I drew the outer suburbs - Prince William, Loudoun, and Manassas. In this second map, I have a more box-like VA-10 with Prince William and Loudoun being entirely separate. The "tail" of the county cluster ends up with the former, the best option. Since VA-10 is just under, and Fairfax+independent cities within its borders is just over two seats, it is natural for a very small share of the latter to be paired with the former. McNair's population meant it was just right for this.

There are eight Biden seats and three Trump seats. However, in 2021, Youngkin won six seats, the median seat actually voting for him by a larger margin than the state as a whole. If I had to guess, it would split 7D-4R, with Sparnbarger still running (and winning) in VA-07 and Wexton narrowly losing VA-10 because of the conservative Shenandoah Valley territory added in. Three districts are majority-minority, two of them in NoVa, though I did not consider the creation of majority-minority seats in how I drew the region.

VA-01 (Newport News): 54W, 33B; 52-46 Biden, EVEN
VA-02 (Virginia Beach): 52W, 31B; 58-40 Biden, D+6
VA-03 (Petersburg): 49W, 39B; 54-45 Biden, D+2
VA-04 (Richmond): 55W, 30B; 62-37 Biden, D+10
VA-05 (Charlottesville): 70W, 19B; 51-47 Trump, R+5
VA-06 (Roanoke): 77W, 13B; 59-39 Trump, R+13
VA-07 (Dale City): 44W, 23H, 22B, 10A; 58-40 Biden, D+6
VA-08 (Alexandria): 46W, 19A, 19H, 16B; 72-26 Biden, D+21
VA-09 (Blacksburg): 83W, 10B; 69-30 Trump, R+22
VA-10 (Leesburg): 63W, 15A, 12H; 51-47 Biden, D+2
VA-11 (Arlington): 53W, 22A; 16H; 74-24 Biden, D+23

DRA link

I have to admit that I find DRA's compactness score quite nonsensical, but if you really find it relevant, then the following map has a score of 56, although it has slightly more county splits. And it doesn't split the Richmond metro and the Hampton Roads core cities three ways:
https://davesredistricting.org/join/20183a8c-ca54-41d3-a30a-76722a45b308
(Just a sketch, not something I really put effort in.)

[President Punxsutawney Phil]

snip

I have to admit that I find DRA's compactness score quite nonsensical, but if you really find it relevant, then the following map has a score of 56, although it has slightly more county splits. And it doesn't split the Richmond metro and the Hampton Roads core cities three ways:
https://davesredistricting.org/join/20183a8c-ca54-41d3-a30a-76722a45b308
(Just a sketch, not something I really put effort in.)

Interesting alternative. The fact it has only one whole-county CD sinks it in my opinion, but a more refined version could be adopted.
Anyway, here the process by which I look at compactness.

I draw districts largely without checking compactness scores. I try to make box shapes for CDs, within reason. The reason why I place some stock in DRA compactness scores is because they seem to have some real heft in telling me if I have done this. Of course. in some places it's just a big fail (like MD, lol). Once I'm done drawing, a bad compactness score could be a warning that I could clean things up a bit. Overall, I think my understanding of compactness is closer to a "KIWYSI" standard, focused a bit more on if any districts look weird and then seeing what it might done about that. It's not guaranteed I take action, especially if I make the judgement that it helps the broader map or my goals to keep things the way they are. County integrity is the most likely reason why I keep things the way they are.

I will say I don't really care too much about metros from a criteria POV. Optimizing compactness over most other factors often means pairing urban jurisdictions with nearby urban or suburban jurisdictions (see: Richmond and Henrico to make a core of a small, compact urban district). Placing a premium on splitting counties and municipalities indirectly guarantees that stopping splits on other things is less important. I don't have problems with splitting metros if needed to reduce splits in other things. This does not mean "you must split metros" is a rule of mine. It just means that I care less if my criteria ends up resulting in that.

I certainly feel as though that given the criteria, it's very hard to pass up drawing not just one but two compact whole-county CDs in Virginia in that area (something I even discovered by accident). Especially since it also leads to a more compact VA-02 and VA-03 as well. It's not like I've never drawn seats that use the St. James river like this either (but then again, I've made many maps). The fact is that the Hampton Roads area is quite big in population. One could draw it cleanly into two seats.

Drawing it the way I've done has some interesting effects of course: it forces VA-05 a bit more north, it drives VA-03 east, and whatever shape VA-04 takes on, it encourages the other seats to flow into wherever the vacuum is. Since on this map it's Chesterfield County, both VA-03 and VA-04 swallow large portions of Chesterfield. One benefit of the Chesterfield split is that, iirc, the black % vs white % in Chesterfield has been increasing, but the opposite is true of the rural parts of VA-03. So this helps keep VA-03 black-performing. I suppose one could shift Norfolk into VA-03, but this would inevitably give VA-03 a "tail" and that seems neither desirable nor necessary.

[President Punxsutawney Phil]

Alabama

75% of the state is covered in whole-county CDs. One of these is the black-plurality AL-07, running almost the entire length of the state. While these are not necessarily compact to a high standard in a general sense, they are compact enough to justify their existence. I tried (and failed) to have no county splits at all. Unfortunately, Jefferson County cannot be paired with any of its neighbors to form a whole-county CD, unlike under 2010 numbers. I divided up the entire area that could not be drawn into whole-county CDs in a way that had compactness in mind and then called it a day. There are five Trump seats and two Biden seats.

AL-01 (Mobile): 63W, 27B; 64-35 Trump, R+16
AL-02 (Auburn): 63W, 28B; 66-33 Trump, R+19
AL-03 (Alabaster): 72W, 20B; 72-27 Trump, R+25
AL-04 (Decatur): 80W; 79-20 Trump, R+32
AL-05 (Huntsville): 69W, 17B; 65-34 Trump, R+19
AL-06 (Birmingham): 50W, 41B; 54-45 Biden, D+2
AL-07 (Montgomery): 47B, 46W; 53-46 Biden, D+2

DRA link

[Libertas Vel Mors]

Alabama

75% of the state is covered in whole-county CDs. One of these is the black-plurality AL-07, running almost the entire length of the state. While these are not necessarily compact to a high standard in a general sense, they are compact enough to justify their existence. I tried (and failed) to have no county splits at all. Unfortunately, Jefferson County cannot be paired with any of its neighbors to form a whole-county CD, unlike under 2010 numbers. I divided up the entire area that could not be drawn into whole-county CDs in a way that had compactness in mind and then called it a day. There are five Trump seats and two Biden seats.

AL-01 (Mobile): 63W, 27B; 64-35 Trump, R+16
AL-02 (Auburn): 63W, 28B; 66-33 Trump, R+19
AL-03 (Alabaster): 72W, 20B; 72-27 Trump, R+25
AL-04 (Decatur): 80W; 79-20 Trump, R+32
AL-05 (Huntsville): 69W, 17B; 65-34 Trump, R+19
AL-06 (Birmingham): 50W, 41B; 54-45 Biden, D+2
AL-07 (Montgomery): 47B, 46W; 53-46 Biden, D+2

DRA link

You can pair Jefferson with Blount if you're willing to take a high pop dev. Jimrtex did that in the Alabama redistricting thread.

[President Punxsutawney Phil]

Alabama

75% of the state is covered in whole-county CDs. One of these is the black-plurality AL-07, running almost the entire length of the state. While these are not necessarily compact to a high standard in a general sense, they are compact enough to justify their existence. I tried (and failed) to have no county splits at all. Unfortunately, Jefferson County cannot be paired with any of its neighbors to form a whole-county CD, unlike under 2010 numbers. I divided up the entire area that could not be drawn into whole-county CDs in a way that had compactness in mind and then called it a day. There are five Trump seats and two Biden seats.

AL-01 (Mobile): 63W, 27B; 64-35 Trump, R+16
AL-02 (Auburn): 63W, 28B; 66-33 Trump, R+19
AL-03 (Alabaster): 72W, 20B; 72-27 Trump, R+25
AL-04 (Decatur): 80W; 79-20 Trump, R+32
AL-05 (Huntsville): 69W, 17B; 65-34 Trump, R+19
AL-06 (Birmingham): 50W, 41B; 54-45 Biden, D+2
AL-07 (Montgomery): 47B, 46W; 53-46 Biden, D+2

DRA link

You can pair Jefferson with Blount if you're willing to take a high pop dev. Jimrtex did that in the Alabama redistricting thread.

How compact could the rest of the districts be? I'm working with a 1% deviation band (i.e. the difference between the biggest district's distance from the quota, and the smallest district, cannot exceed 1%). Under this, -0.59% is okay is there is no district above +0.40%, but if it becomes +0.44%, then something has to change on either the lower end or the higher one.

[President Punxsutawney Phil]

Kansas

I tried to draw a whole-county CD involving Johnson County. My first candidate was Lawrence-Johnson, 6k off. In the end it was not to be; I was not interested in having a one-county-wide KS-03 running along the Missouri border. Instead I went with as compact a KS-03 as I could while splitting Johnson County. I still was committed to no county splits elsewhere. The rest of the map effectively drew itself. I ended up with 3 box-like CDs to my satisfaction. There is one Biden district and three Trump districts.

KS-01 (Manhattan): 74W, 17H; 70-28 Trump, R+25
KS-02 (Topeka): 79W, 56-42 Trump, R+10
KS-03 (Overland Park): 67W, 14H, 10B; 56-42 Biden, D+3
KS-04 (Wichita): 70W, 13H; 60-38 Trump, R+14

DRA link

[President Punxsutawney Phil]

All states in the Midwest and South now have maps. Three states left: MT, HI, and CA.

Midwest: 10/10 (37/54)
Northeast: 8/8 (54/21)
South: 15/15 (74/90)
West: 8/11 (24/24)
Nationwide: 41/44 (189/189)

HI and MT both have 2 seats and Ds and Rs are likely to sweep respectively. Unless a changing of maps somewhere happens (far from impossible, depends), the nation less CA will be a tie.

[President Punxsutawney Phil]

Montana

This was the obvious first option. Not a lot to see here. I was always going to avoid splitting any counties if I could feasibly do so.

MT-01 (Missoula): 87W; 53-44 Trump, R+8
MT-02 (Billings): 80W, 13N; 61-36 Trump, R+15

DRA link

[President Punxsutawney Phil]

Hawaii

There wasn't much to do here. I simply sought to get rid of any municipal splits and draw a compact HI-01. Don't let the colors fool you...I changed the district numbers after I drew it to march them to the current ones. Needless to say, neither of these seats are feasibly voting for Trump in 2024.

HI-01 (Honolulu): 65A, 22P, 17W; 65-34 Biden, D+15
HI-02 (Hilo): 49A, 32P, 26W, 11H; 63-35 Biden, D+14

DRA link