Dave's Redistricting App

[Verily]

Democratic gerrymander of PA. Guarantees the defeat of Jim Gerlach and makes Charlie Dent likely to lose while also preserving all incumbents except Critz, who is doomed.

[Verily]

Yes; PA-2 is 55% black (could be more easily).

PA-1: 62-37 Obama
PA-2: 81-19 Obama
PA-3: 56-43 Obama
PA-4: 59-40 Obama
PA-5: 36-63 McCain
PA-6: 63-36 Obama
PA-7: 65-34 Obama
PA-8: 60-39 Obama
PA-9: 37-62 McCain
PA-10: 49-50 McCain (the best one can do for Carney, but he should be reasonably safe)
PA-11: 56-43 Obama
PA-12: 41-58 McCain
PA-13: 62-37 Obama
PA-14: 63-37 Obama
PA-15: 58-41 Obama
PA-16: 36-63 McCain
PA-17: 55-44 Obama
PA-18: 38-61 McCain

[Verily]

Critz probably could win that PA-12 seat, it's hardly a Republican seat so much as an anti-Obama one. Kerry probably got at least around 45% there.

Unfortunately for Critz, he lives in Johnstown, which is split between PA-3 and PA-9. And there's no way he could win PA-12 anyway.


I'm working on a Republican map but have run into difficulty creating only 3 Obama seats in SE PA. It may be impossible.

[Verily]

You also put Holden in PA-10, and he'd probably knock off Carney in a primary. Dent might very well run there too.

Interesting. I assumed Holden lived around Harrisburg. But Carney would surely win a Carney-Holden primary in PA-10; Carney represents a lot more of the district than Holden does.

Dent couldn't beat either Holden or Carney in that PA-10. The district doesn't fit his style, and he barely represents any of it right now.

[Verily]

Actually, I have now managed to create just three Obama seats in SE PA, plus another Obama seat containing Democratic parts of Bucks County, the most Democratic parts of the Lehigh Valley, and the Scranton/Wilkes-Barre area. This allows for PA-6, PA-7, PA-8 and PA-15 to all become McCain seats, albeit not particularly strongly (still much better than the current arrangement for Republicans). PA-17 is more Democratic, but it loses Schuylkill County and should be unwinnable for Democrats not named Tim Holden.

The map's not quite finished, though.

[Verily]

That equals five Obama seats total I assume since there will be one more in Pittsburgh?

There are six total. Four in eastern PA and two in western PA.

Eastern PA:
PA-1 (plurality black Philly and SE Delaware County)
PA-2 (majority black Philly and a bit of Delaware County)
PA-11 (Scranton/Wilkes-Barre, central Lehigh Valley, E Bucks County)
PA-13 (snaking through Delaware, Chester and Montgomery Counties as well as parts of Philly)

Western PA:
PA-4 (Erie, Ohio River Valley, N Pittsburgh)
PA-14 (Pittsburgh and Alleghany Valley)

I chose to create two solidly Democratic seats in Western PA to guarantee the defeat of two of Critz, Altmire and Dahlkemper; getting rid of all three risked letting all three hang on.

[Verily]

How does one know how Hispanics actually voted?  Exit polls?  Or is the game just one involving what surname seems to get elected? Has SCOTUS ruled on this notion, that if 50% +1 Hispanic VAP seems to still elect Anglos, and in particular Anglo Republicans, that further inquiry is needed?

What I'm saying is that SCOTUS has not ruled on this specific question. However, they have clearly left open the possibility that the state could find a way to make a 50% VAP district that would be drawn to favor the white candidate instead of the minority candidate. The De Grandy court has intentionally left itself the ability to rule against the state in that circumstance.

OK. So if Bonilla had won anyway, due to the remaining hyper GOP Anglo precincts remaining in that CD after being redrawn, we might have had another case! Will we have another case if Ciro loses?  Sure his opponent is Hispanic, but Ciro will get the majority of the Hispanic vote, or of the fictive Hispanic VAP vote (if they all in some alternative universe assuming citizenship voted in percentage numbers equaling the Anglos) in all events. I guess not given the place we are at on the decennial calendar. This is just so much fun, not.

Not if it were a one-off election; LA-02 is not a VRA violation. But if Rodriguez lost, and his Republican replacement were consistently reelected without winning the Hispanic vote, then it would be a VRA violation until and unless the replacement began winning a majority of the Hispanic vote (a la Steve Cohen).

[Verily]

Just to show the rapidness of the political transformation... Clinton won Clayton County, GA 45-41 in 1992. By 2000, Gore won it 65-33. And in 2008, Obama won it 83-17. Similar is now happening in Henry, Newton and Rockdale Counties, and also in nearby Douglas, Cobb and Gwinnett Counties.

[Verily]

New Hampshire? Why?

Also interesting that the Jamaica district is not VRA-protected. Better gerrymandering in East Queens, here we come.

[Verily]

CDs 4 and 5, too. The only place you should be crossing the Cascades is along the Columbia River at the southern edge of the state. And Island and San Juan Counties are only accessible from the eastern side of Puget Sound; connecting them to the western side makes no sense at all.

CD-8 is very gerrymandered; it's blatant gerrymandering to include two unrelated cities and nothing else in a district.

Basically, the whole map is a Republican gerrymander.

[Verily]

There's no need to make the map so complicated. Have CO-05 take in all the ski lands and wrap around the west of Denver instead of the east. Put Greeley into the Boulder district to make up for the lost population. There should be a map like that that I did recently lying around somewhere...

[Verily]

Greeley isn't a Democratic city though. You can't see it well on the map that far out, but I had the yellow district take in only the eastern half of it (the older part of the city which is also heavily Hispanic.) The western half (which looks like new development and suburban hell on Google Street View) is in the gray district.

Greeley's not particularly Republican (maybe the city itself even voted for Obama), and it doesn't matter anyway. You'd have to be putting Boulder with something like Grand Junction for a district containing it to be at all vulnerable.

[Verily]

Remember the district already contains southern Jefferson County and the most densely populated part of Douglas, Tom Tancredo territory. It can easily absorb and handle that now, but something 50/50 like Greeley might leave it a bit too open.

Not really. Try this (CO-01 is majority Hispanic):

[Verily]

Here's a deceptive gerrymander of NM. Every district is 57% for Obama, but only one county (Bernalillo) is split. NM-01 and NM-02 are majority Hispanic, while NM-03 is 41% white, 34% Hispanic, 21% Native. (All three districts are either 41% or 42% white as well.)

It's not obvious from this screenshot, but Las Cruces is in NM-01.

[Verily]

I am, too.

[Verily]

It started working again this evening for me.

[Verily]

The brown district is 51% white, 43% black. The borders didn't actually change much, now that I look at it; Brooks, Dooly, Crisp, and Lowndes Counties got removed, and Peach, Harris, and parts of Muscogee and Colquitt got added.

It's possible to make it majority black. Although it isn't majority now, after his scare I think Bishop may ask the Obama administration/DOJ to pressure Georgia to make it so. And I think the Georgia Republicans wouldn't mind shoring up the mostly unbeatable Bishop to ensure that no new Dixiecrats pop up in South Georgia.

[Verily]

Weird, they must hold some pretty conservative areas then. Some of those districts would've had to voted for O'Donnell. I have a tough time seeing how anyone could vote for both O'Donnell and a Democrat.

Knowing rural Delaware, at least some of those Democratic legislators from Sussex County probably voted for O'Donnell.

[Verily]

Weird, they must hold some pretty conservative areas then. Some of those districts would've had to voted for O'Donnell. I have a tough time seeing how anyone could vote for both O'Donnell and a Democrat.

Knowing rural Delaware, at least some of those Democratic legislators from Sussex County probably voted for O'Donnell.

Delaware has ConservaDems? Kind of weird, it's traditionally Republican rather than Democratic and never had any type of Dixiecrat machine. FDR even failed to win it in 1932.

Well, yeah, New Castle County has always dominated the state, and back when Philly was GOP, so was Wilmington. But Sussex and to a lesser extent Kent County might as well be in rural Alabama, politically. See also: Eastern Shore of Maryland.

[Verily]

I'm curious if anyone knows--has Dave's Redistricting been updated for the 2009 ACS figures? I ask specifically about Georgia (I know he has new figures for New England.) I was poking around the NYT map, and then verified this info on the American FactFinder Census site, and the black population is increasing so fast in some parts of the suburban Atlanta metro that it may in fact be not only possible but easy to draw a fourth black majority seat in the Atlanta metro (which could affect whether the DOJ  sues to force one).

For example, Rockdale County, GA was estimated at 32.6% black in 2008 ACS and is now estimated at 37.0% black in the 2009 ACS. On that 3.4% annual increase, it would be around 39% black at the 2010 Census (which is less than a full year after the ACS). Similar rapid growth is shown in Newton County, Henry County, Douglas County and Cobb County (and to a lesser extent in Gwinnett County, where Hispanics and Asians form the larger part of the new residents, and in Clayton County, which increased from 51% black in 2000 to 60% in the 2009 ACS but seems to have stabilized recently). It is accompanied by a slight decline in Fulton County, but the decline in Fulton is tiny relative to the increase in the others. If the changes really are that rapid, how easy a black district is to create might change dramatically from the 2008 estimates to the 2009 ones, or to the actual 2010 census figures (which will include another half-year of change after the 2009 estimates). And even on the 2008 numbers it is possible, just barely, to craft four black majority seats in the area (albeit only 50-51% black each).

[Verily]

Just checked, and it appears Georgia at least is up-to-date on the block group version. Awesome; time to try to draw 4 black majority seats in the Atlanta metro.

[Verily]

- Obama Hispanic district in Florida

Already exists. Ileana Ros-Lehtinen's seat voted for Obama. You could probably make two if you diluted the Cuban vote more (but the data isn't out yet).

[Verily]

As for West Virginia, I doubt it.  his best county was Boone, with 54% of the Vote, and there's no county in WV big enough where you might be able to draw an Obama-leaning part out of it.

Sure there is. Kanawha County has very clear polarization between the poor, Democratic areas along the river and the less poor, Republican areas uphill. I'm sure you could draw an Obama seat in WV if you had partisan data by voting district.

[Verily]

Yeah, Texas racial numbers are definitely off since the partisan data was added (maybe the partisan ones, too, didn't check). It looks like the precinct data got randomized within each county, which means its only really obvious in the large, diverse counties. But it's absolutely wrong.

[Verily]

The Georgia data has been updated to reflect the absentee/early voting totals.

A few precincts, especially around Columbus, seem to be missing all of the data.