Dave's Redistricting App

[JohnnyLongtorso]

For what it's worth, the district I made is about 48% black.

Dave's posted on SSP about how to help get more partisan data on the app:

http://swingstateproject.com/diary/7420/redistricting-app-how-you-can-help-get-partisan-data

Edit: for those that don't care to read the article, he also mentions New Mexico has partisan data now.

[JohnnyLongtorso]

California, Maryland, New Mexico, New York, North Carolina, and Texas. You have to push the "test data" button first for New York.

[JohnnyLongtorso]

When did New Mexico get partisan data?

Today.

[JohnnyLongtorso]

Hey guys, Dave's added partisan data for Pennsylvania.

[JohnnyLongtorso]

And now to check and see how I did here...

Here's a Republican map of Pennsylvania, losing a district from the current map. Tried not to overreach too much; concentrated on protecting incumbents and watering down a couple of districts, rather than trying to maximize Republican seats.



Click for bigger.

PA-01 (blue Philly district, Bob Brady - D) - Expands to take in more of Delaware County, uber-safe for the Dems.
PA-02 (green Philly district, Chaka Fattah - D) - Mostly unchanged; still majority-black. Obviously uber-safe.
PA-03 (purple NW corner district, Kathy Dahlkemper - D) - Pushes east instead of south, taking in some heavily-Republican counties. Should make the district more Republican.
PA-04 (red SW district, Jason Altmire - D) - Takes in all of the Dem-leaning areas outside of Pittsburgh. Should be safe for Altmire.
PA-05 (orange central district, Glenn Thompson - R) - Takes in a bunch of the Philly suburbs in Westmoreland County instead of the NW counties, which shouldn't change the partisan balance much.
PA-06 (teal SE PA district, Jim Gerlach - R) - Tried to make this district as safe as possible, so it takes in all the Republican parts of Montgomery and Bucks now. Should be a Republican-leaning district now.
PA-07 (grey SE PA district, open) - Pushed west in order to make it less Democratic. Still a swing district, but should be more favorable to the Republicans.
PA-08 (purple SE PA district, Patrick Murphy - D) - Takes in all the Dem parts of Bucks, and parts of Montgomery and Philadelphia. Safe Dem seat.
PA-09 (light blue SW PA district, Bud Shuster - R and Mark Critz - D) - Swallows parts of Critz's district, which tilts the district from "overwhelmingly Republican" to just "very Republican". Shuster would prevail in a matchup with Critz.
PA-10 (magenta NE PA district, Chris Carney - D) - Carney's district is completely reconfigured as a safe Dem district, taking in all of Lackawanna along with Bethlehem, Scranton, and part of Allentown.
PA-11 (green NE PA district, Paul Kanjorski - D) - Turns Kanjo's district (since he's probably going to lose to Barletta) into a pretty Republican one. Shouldn't be hard for Barletta to hold in 2012.
PA-12 (light purple SE PA district, Todd Platts - R) - Takes in Harrisburg, which makes the district a little less Republican, but it shouldn't be hard for Republicans to hold.
PA-13 (pink SE PA district, Alyson Schwartz - D) - Heavily-Democratic MontCo-based district.
PA-14 (brown SW PA district, Mike Doyle - D) - Heavily-Democratic Pittsburgh district.
PA-15 (orange NE PA district, Charlie Dent - R) - With Bethlehem and most of Allentown gone, the district picks up some Republican-leaning counties that should make it easier for Dent to hold.
PA-16 (green SE PA district, Joe Pitts - R) - Mostly Lancaster County, takes in some Dem parts of Berks and Chester, but should remain Republican-leaning.
PA-17 (dark purple SE PA district, Tim Holden - D) - Getting rid of Tim Holden requires chopping up Schuylkill County, and also removes Harrisburg. Should be even more Republican now.
PA-18 (yellow SW PA district, Tim Murphy - R) - Takes in a bunch of Republican-leaning parts of SW PA. Maybe a little less Republican, but still Republican. Republican!

PA-01 - 85-14 Obama (from 88-12 Obama)
PA-02 - 87-13 Obama (from 90-10 Obama)
PA-03 - 50-49 Obama (from 49-49 McCain)
PA-04 - 51-48 Obama (from 55-45 McCain)
PA-05 - 54-45 McCain (from 55-44 McCain)
PA-06 - 49-49 Obama (from 58-41 Obama) - Obama won this district by 343 votes.
PA-07 - 52-47 Obama (from 56-43 Obama)
PA-08 - 57-42 Obama (from 54-45 Obama)
PA-09 - 60-38 McCain (from 63-35 McCain)
PA-10 - 63-36 Obama (from 54-45 McCain)
PA-11 - 55-43 McCain (from 57-42 Obama)
PA-12 - 53-46 McCain (from PA-19's 56-43 McCain)
PA-13 - 63-36 Obama (from 59-41 Obama)
PA-14 - 66-33 Obama (from 70-29 Obama)
PA-15 - 50-48 Obama (from 56-43 Obama)
PA-16 - 50-49 McCain (from 51-48 McCain)
PA-17 - 58-41 McCain (from 51-48 McCain)
PA-18 - 58-41 McCain (from 55-44 McCain)

Overall, not bad, although I didn't help Republicans enough in PA-03.

[JohnnyLongtorso]

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

[JohnnyLongtorso]

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.

He's a lifelong resident of Schuylkill County, and was the sheriff there before he was elected to Congress. It remains a ridiculously loyal base of support for him.

[JohnnyLongtorso]

Inspired by the brief discussion of Oregon's redistricting process (which prevents party affiliation as a basis for drawing districts, and protects communities of interest), I thought I'd try my hand at a nonpartisan map of the state:



I'm not sure if splitting up Portland and Gresham would be kosher, but otherwise I think the map is pretty solid. The blue, purple, and red districts would all be swing districts, with the Dems probably having a slight advantage in the blue and red ones.

[JohnnyLongtorso]

With the midterm result in, here's an idea for a bipartisan incumbent protection map of Virginia:



Click for huge.

VA-01 (Rob Wittman - R) - Hasn't changed a whole lot; takes in some more of Prince William, loses part of Stafford/Spotsylvania, and adds the two Eastern Shore counties (they were in VA-01 prior to the 2000 map).
VA-02 (Scott Rigell - R) - Snakes up the shore to take in some Republican territory (Poquoson, parts of York and Gloucester), picks up the VA-01 bit of Hampton, and loses the Eastern Shore. Should move the needle to the Republicans by a couple points.
VA-03 (Bobby Scott - D) - Takes Petersburg out of VA-04, mostly unchanged. 62% black.
VA-04 (Randy Forbes - R) - Also not changed a whole lot, aside from losing Petersburg, which should flip the district to McCain.
VA-05 (Robert Hurt - R) - Removes Charlottesville and most of Albemarle, adds the rest of Bedford, Lynchburg, and Amherst. Should be no trouble at all for the Republicans to hold now.
VA-06 (Bob Goodlatte - R) - Snakes up from Roanoke, where Goodlatte lives, through the Shenandoah Valley, and pulls in Charlottesville and Albemarle and some outer NoVa counties. Shouldn't endanger Goodlatte.
VA-07 (Eric Cantor - R) - Actually a little less ridiculous now, it's a solidly-Republican suburban Richmond/Fredericksburg area district. Still should be solidly Republican.
VA-08 (Jim Moran - D) - Remains solidly Dem; continues to hold the trifecta of Alexandria, Arlington, and Falls Church, but goes west instead of south. Maybe a couple points less Democratic, but still a safe D district.
VA-09 (Morgan Griffith - R) - Not much changed here; added Salem, Martinsville, and some more of Roanoke County. Safe R.
VA-10 (Frank Wolf - R) - Added almost all of Shenandoah County, and removes some parts of Fairfax. Retains the most Republican parts of Fairfax. Safe for Wolf, should lean Republican in an open seat, unless it's a particularly good Dem year.
VA-11 (Gerry Connolly - D) - Replaces the Republican PW County parts with the solidly-Dem SE PW County. Should be a pretty safe Dem seat.

[JohnnyLongtorso]

I've been playing around with making "nonpartisan" maps. My general standards are that the districts should be relatively compact without dividing up counties/cities, and partisanship isn't taken into consideration. I do preserve existing VRA districts, though.

Here's one for Georgia (14 districts):



The green, purple, and blue districts in the Atlanta area are all 51% black. It would probably maintain the status quo; four of the five current Dems have pretty much the same maps, and Barrow would likely run in the light green district, which is 39% black. The teal Gwinnett County district might be interesting in a decade or so; according to the app it's only 48% white, but of course, the suburbs are still pretty solidly Republican for now.

And here's one for Michigan (also 14 districts):



The two majority-black seats retain their majorities. Democrats would easily win those, the red Oakland, purple Wayne/Oakland, and gray Flint/Saginaw districts. The yellow Macomb district, green Lansing-area district, teal Washtenaw, and possibly even the light blue midstate district would be swing districts. Republicans wouldn't have much trouble holding the remaining five districts.

[JohnnyLongtorso]

Dave is supposed to be releasing a new version of the app today. All the old maps should still work with it, but it should have some improved features, I guess.

That said, here's another nonpartisan map: Ohio (16 districts).



I managed to (just barely) preserve the black majority district without it plunging into Akron. On this one, the Dems would have five safe districts: the blue, green, red, and yellow NE ones, the new purple Columbus district. Republicans would have no trouble holding the magenta district, light blue Columbus suburban district, light purple and brown Cincinnati suburban districts, light green NW district, and probably the teal Erie to Holmes district. In addition, the pink Dayton district would probably favor Republicans, although not overwhelmingly so. Similarly, the orange Toledo-area district would still lean to the Democrats, but not as strongly as before. That leaves the Cincinnati district, which would remain a swing district, the purple NE district, which gets pushed towards the Dems but also remains swingy, and the grey mid-eastern district, which would probably lean to the Republicans but not by a large margin. So, a 7R-6D map with three swing districts.

[JohnnyLongtorso]

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.

[JohnnyLongtorso]

Version 2.0 of the app has been released. For those of you with bookmarks directly to the app, you'll need to change it to this URL:

http://gardow.com/davebradlee/redistricting/davesredistricting2.0.aspx

It's certainly an interesting change, but it'll take some getting used to.

[JohnnyLongtorso]

It's hard to say what the numbers for the Cincinnati district would be, since there's no precinct-level breakdown handy. I guess, on second thought, it would probably lean to the Dems.

The magenta district, eyeballing it, would probably have voted in the mid-50s for McCain. No problem for a Republican who's not Jean Schmidt to hold, and even Schmidt has quieted down (for the most part).

From a competitiveness standpoint, a Dayton/Springfield district would be more logical, but it would go against my nonpartisan, "communities of interest" approach, since presumably western Greene County is suburban Dayton.

[JohnnyLongtorso]

I find the new version much easier to use when coming up with nonpartisan maps; being able to see the actual lay of the land makes the "communities of interest" standard easier to apply.

[JohnnyLongtorso]

In celebration of this achievement, I've made 2-CD maps of all the at-large states (minus Alaska), with the intent of giving the minority party in each state a CD they could win.

Delaware:



The green district might have actually been a narrow Obama win, but it would be dominated by Kent and Sussex, so Republicans would be able to win it pretty easily. The blue district is probably over 70% Obama.

Montana:



The blue district is around 53-44 Obama, while the green one is around 55-42 McCain.

North Dakota:



This one, along with Delaware, has the advantage of looking pretty nice as well. No county splits even! The blue district is about 52-46 Obama, while the green one is about 60-38 McCain.

South Dakota:



Less so here. Blue district is pretty much even, while the green one is about 58-40 McCain.

Vermont:



Really no way to make a Republican-opportunity district here. The blue district almost certainly is over 60% Obama, while the green one is over 70% Obama.

Wyoming:



Opposite problem, thought not *quite* as bad. My guess is the blue district is around 60% McCain, while the green one is over 70%.

[JohnnyLongtorso]

In the Senate, they've got SD-14 (split between New Castle and Kent) SD-17 (Dover), SD-20 (beach towns, Millsboro, Selbyville), and SD-21 (Laurel/Delmar).

SD-17 and 20 were up last year, the former being a 2-1 win and the latter unopposed. They lost SD-15 last year, which is the western half of Kent County.

In the House, there's HD-14 (Rehoboth), HD-28 (Smyrna), HD-31 (Dover), HD-32 (Dover), and HD-41 (Millsboro), though HD-41's representative is only a Democrat because he was a Republican who had to resign from the House due to a scandal (domestic abuse, to be specific), then he came back and ran as a Democrat.

The 14th was close (54-46) last year, 28th and 31st weren't, the 32nd was a 50-44 win, and the 41st was a 55-45 win. They lost the 33rd 52-48 last year, which is Milford and Riverview.

Here are the current maps:

http://elections.delaware.gov/information/districtmaps/districtmaps.shtml

[JohnnyLongtorso]

Results by House district. Coons carried HD-14 by 49-48 (about 150 votes), HD-28 by 55-42, HD-31 by 57-39, and HD-32 by 51-44.

Looks like HD-41 was the one Dem-held House district that voted for O'Donnell, 58-38. HD-33, which the Republicans picked up, went for O'Donnell as well (57-39). Meanwhile, Coons carried every seat in New Castle County, including the five Republicans left there (HD-11, 12, 20, 21, 22), and the three House seats they picked up (HD-6, 9, and 24).

After a quick calculation of the Senate districts that were up last year... SD-14 went for Coons by 55-41, while SD-20 went for O'Donnell 56-40. SD-17, which the Dems lost, went for O'Donnell 54-42.

So that's two districts that were up that went for O'Donnell/a Democrat. As I've said, Atkins is a DINO, and Bunting (the Senator from SD-20), aside from being unopposed, is pretty conservative, I believe.

[JohnnyLongtorso]

Yeah, Sussex and Kent culturally belong with the Eastern Shore of Maryland. You could cut the state off at the Chesapeake and Delaware Canal and put the southern part in with Maryland and I doubt anyone would complain.

Sussex actually still has a slight Dem registration edge (52k to 49k), while Kent's is more pronounced (46k to 33k).

[JohnnyLongtorso]

Still kind of odd that they'd still be around this late, and that the guy in HD-41 was even welcomed into the caucus (I'd be pretty outraged about that if I were a Delaware Democrat, though I don't know the whole story, maybe they were pretty reluctant.)

I just found something odd though it might just be an error, there's a precinct near Wilmington which is literally 100% white, and it has almost 1000 people (close to Canby Park). It's about in the area where the black population starts to fade, but not to that level, the surrounding precincts are about 20% black.

Democrats didn't pick up the House until 2008, so they may have been hedging their bets in case Atkins became the deciding vote for Speaker. A little googling suggests that he presented himself as a "changed man", so maybe he's cleaned up his act.

That precinct is probably just an error. I looked at the neighborhood on Google and it's a mix of townhouses, split-levels, and other small single-family homes; doubtful that it's 100% white.

Edit: Although maybe not, there are some other heavily-white precincts in the area (near the cemetary). Maybe it's a white working-class neighborhood like Hampden in Baltimore.

[JohnnyLongtorso]

He's been updating the states with the ACS figures, but I'm assuming he hasn't gotten them all in yet. When he does updates, he posts them on SSP and on the launch page:

http://gardow.com/davebradlee/redistricting/launchapp.html

[JohnnyLongtorso]

I posted a diary about redistricting the Virginia Senate at SSP. Check it out if you like.

[JohnnyLongtorso]

The app has been updated with the 2010 census data and 2008 partisan data for Virginia.

[JohnnyLongtorso]

White-majority Obama district in NM:



It's easy to make one by connecting Santa Fe with eastern Bernalillo and Los Alamos, so I tried to make one that was at least 60% white.

And here's a cornucopia of white-majority Obama districts in Texas (I didn't use the new pop estimates because they apparently reduce the total population for some reason; I did use the new apportionment numbers):



You may have to open the latter in a new window to see it better.

[JohnnyLongtorso]

The app has been updated with 2010 census data for all the states that are currently out, with the exception of Oregon.