Need super-programmer to help calculate redistricting

[GunnyFreedom]

OK, this is a very complicated problem. It's complicated enough to redraw the districts in a way that obeys State law while attempting to balance the partisanship towards zero. Add in the Federal Voting Rights Act, and it gets ridiculously complicated. The best idea I have at this point is to use a computer to crunch the millions of potential permutations until the most fair, legal, and balanced districts are created.

Here are the parameters:

1) The NC State Constitution says that counties cannot be split.

2) Federal Voting Rights Act pretty much requires at least 2 majority-minority districts, thus requiring that we violate #1, but do so as little as possible.

3) Voting Districts need to be as "compact" as possible, so no long snaking thin corridors connecting sections, or (given #2) as little as possible.

4) Population within districts needs to be as even as possible.

5) My personal preference is to make districts as competitive as possible, balancing partisanship towards zero. That will be impossible in the two districts required by #2 above.

NC will certainly have 13 congressional districts an possibly 14, so maps need to be produced for both eventualities.

NC will also be redrawing the 120 State House Districts and the 60 State Senate Districts, so maps will need to be produced for legislative redistricting as well.

MY THOUGHTS:

We could just use the results from the 2008 Presidential election to calculate partisanship, but to do so would clearly favor Republicans during redistricting. I would prefer to get precinct-by-precinct data for the last 5 elections and average them per precinct by % between Presidential, Gubernatorial, and Senate elections, and then average them per precinct across all 5 years. That would provide the most "fair" picture of precinct partisanship, and it should not be that difficult to produce. I am sure I can call on the SBOE and get them to provide that data to me as a State House Rep. A couple hours worth of spreadsheet work could produce the detailed 5 year partisanship data, once we have the base data to work with. Apply the resulting % in each precinct to precinct population, and you have the person-by-person data necessary for balancing the districts.

Once partisan data is plugged into the precincts, the problem gets complicated and will require the computer. The program in question will need to group counties by proximity, precincts by county, as well as by precincts by proximity.

My guess is that the two VRA majority-minority districts will need to be drawn by hand -- too many variables to program easily once you consider that almost every aspect of those two districts will be built on "exceptions" from the primary algorithms. If we were to try and get the computer to map those as well, all VRA district precincts need to share linear borders, be contiguous, and less than 50% white.

All districts will need to be composed of counties of precincts that share linear borders and are contiguous.

Myself and another member of RPF's are currently working to produce fair and legal models by hand, but the balancing act is incredibly tough.

So I am calling for help from a major mondo super computer programmer to help set up a program that can analyze the data and produce a map of voting districts that include the 2 VRA districts, and then balance the remaining precincts as evenly a possible by population, and partisanship, maintaining contiguous counties as much as possible.

I can see possible ways to do it, but writing the program is way beyong my capabilities.

Once we have the partisanship data by precinct, and we remove the two VRA precincts from the mix, the remaining precincts will yield a partisanship of X%/Y% (say 48%/52%). Every single remaining district then should end up as close to 48/52 as mathematically possible given the parameters mentioned above. That way, partisanship is dispersed as evenly as conceivably possible.

So please help -- hard core programmers!

Do not underestimate the critical importance of this effort. Your help will help BJ Lawson win. Your help will help Glen Bradley win when he decides to run for Congress. This effort could directly put two more Ron Paul's into Congress.

If you are unable to do this, but know someone who can, PLEASE we really need your help!

thanks!

[newbitech]

I always find these types of problems interesting. I have a copy of Micro Station which I think would be a map drawing program that could handle this solution. Micro Station can be programmed with Visual Basic, C#, or the built in scripting language.

There are off the self solutions such as ArcView10 from ESRI that will do exactly what you want it to do with all of your parameters tweaked inside the options of the system.

I haven't seen ArcView10 at work, but I imagine it to be a vector graphics drawing tool that has solved all of the complex problems that would need to be solved in a custom Micro Station solution. These problems will need to be solved before applying your specific parameters.

I am not sure if you have a budget for this project, but at 1,500 dollars for the ArcView10 "template", you'd be able to spend your time focusing on getting the parameters perfect, rather than trying to do things like figure out how to connect to a database where your geo-data will be stored. You won't have to write the data queries or the reports etc..

[MozoVote]

Sounds like what you're proposing is something like what the "Swing State" was ostensibly supposed to be. Although most of the posters there are proud of their progressive-favoring maps.

One thing to decide on early on, is what constitutes VRA compliance. Is over 50% "non white" necessary? There is some ambiguity in what the Justice Dept wants - if the new districts are geographically compact and not demonstrably being done to deny minority votes, maybe as little as 40% is sufficient.

For North Carolina alone - you probably don't need a computer, so much as several nerds willing to huddle over maps and paint census tracts. But if you wanted something that other people could pick up and run with in other states, it could be helpful to them as a starting point.

[GunnyFreedom]

I always find these types of problems interesting. I have a copy of Micro Station which I think would be a map drawing program that could handle this solution. Micro Station can be programmed with Visual Basic, C#, or the built in scripting language.

There are off the self solutions such as ArcView10 from ESRI that will do exactly what you want it to do with all of your parameters tweaked inside the options of the system.

I haven't seen ArcView10 at work, but I imagine it to be a vector graphics drawing tool that has solved all of the complex problems that would need to be solved in a custom Micro Station solution. These problems will need to be solved before applying your specific parameters.

I am not sure if you have a budget for this project, but at 1,500 dollars for the ArcView10 "template", you'd be able to spend your time focusing on getting the parameters perfect, rather than trying to do things like figure out how to connect to a database where your geo-data will be stored. You won't have to write the data queries or the reports etc..

I may be able to come up with $1500, but I'm not sure about the timeframe. If I get on the committee for redistricting (possible) those funds may actually be allocated. Otherwise I may be able to raise the funds and spend it out of my reelection warchest. In any case, we'll still need someone of expertise to operate the application.

[MozoVote]

You may need to look into the legal technicalities if you raise money. Get some people from both parties and a few independents involved, so that you can argue it's non-partisan. A separate N.C. committee account would be a good idea.

I remember the NC League of Women Voters proposing a map of their own in the 1990s. I think it got shot down, though.

[GunnyFreedom]

Sounds like what you're proposing is something like what the "Swing State" was ostensibly supposed to be. Although most of the posters there are proud of their progressive-favoring maps.

One thing to decide on early on, is what constitutes VRA compliance. Is over 50% "non white" necessary? There is some ambiguity in what the Justice Dept wants - if the new districts are geographically compact and not demonstrably being done to deny minority votes, maybe as little as 40% is sufficient.

For North Carolina alone - you probably don't need a computer, so much as several nerds willing to huddle over maps and paint census tracts. But if you wanted something that other people could pick up and run with in other states, it could be helpful to them as a starting point.

Well, the idea here is to create a methodology that is as much as possible immune from gerrymandering so that the precedent extends into the next Democrat-majority legislature. Therefore my goal will be to not only have the computer program balancing the districts, but further a set of guidelines established for operating the program that will be as fair as possible no matter who operates the redistricting.

I'm trying to take the long view. The GOP won't always be in charge of redistricting, and while I believe that we will be significantly more fair than the Democrats have been these last 120 years, if we don't put some methodology in place which assures fairness for BOTH parties, then we'll eventually end up right back where we started -- whether that means in 2020 the Dems regerrymander it to Dems or the Reps regerrymander it to Reps. I don't like either idea.

This year 2010/2011 even with Reps in charge it is likely to end up fair and legal just because of the exasperation we have felt for over a century people want to demonstrate that "we are not like those crooks," but leave the Reps alone for 10 years and in 2020 we may end up actually tilting the balance the other way.

So instead I am thinking about setting up a methodology and guidelines that will to the best of our human ability, use an impartial computer to produce as impartial a map as possible given the counties and parameters we have to work with. Then, if at all possible, push Skip Stam's earlier "non partisan redistricting commission" to operate said methodology to put the final nail into the coffin and guarantee fair voting districts for decades and decades to come.

[Rael]

Will the legislature get rid of mel watts gerrymandered district?

[driege]

Here are the parameters:

1) The NC State Constitution says that counties cannot be split.

2) Federal Voting Rights Act pretty much requires at least 2 majority-minority districts, thus requiring that we violate #1, but do so as little as possible.


Are you sure the Federal Voting Rights Act requires majority-minority districts. I'm sure you would agree this is a bad thing and I don't think you are required to do it.

According to Wikipedia: "The United States Supreme Court in Miller v. Johnson, 515 U.S. 900 (1995), overturned a 1992 Congressional redistricting plan which had created minority majority districts in Georgia as unconstitutional gerrymander. In Bush v. Vera, the Supreme Court, in a plurality opinion, rejected Texas's contention that Section 5 required racially-gerrymandered districts."

Link: http://en.wikipedia.org/wiki/Voting_Rights_Act#Gerrymandering

[nate895]

You don't necessarily need one of us to do this. Most states actually have redistricting programs available to legislatures.

[driege]

OK, this is a very complicated problem. It's complicated enough to redraw the districts in a way that obeys State law while attempting to balance the partisanship towards zero. Add in the Federal Voting Rights Act, and it gets ridiculously complicated. The best idea I have at this point is to use a computer to crunch the millions of potential permutations until the most fair, legal, and balanced districts are created.


Gunny, rather than trying to create balanced districts, which does have seem like a positive change from the status quo, I tend to think districts should be drawn without regard for the racial or political demographics (black vs white, Dem vs Rep). Rather, they should be drawn as the simplest polygons possible with equal numbers of constituents in each. What is your opinion of this?

[nate895]

Gunny, rather than trying to create balanced districts, which does have seem like a positive change from the status quo, I tend to think districts should be drawn without regard for the racial or political demographics (black vs white, Dem vs Rep). Rather, they should be drawn as the simplest polygons possible with equal numbers of constituents in each. What is your opinion of this?

Regulations demand discrimination on the basis of race in states that have enough minorities to create "minority-influence" and "majority-minority" districts.

[driege]

Regulations demand discrimination on the basis of race in states that have enough minorities to create "minority-influence" and "majority-minority" districts.

See my post above. It doesn't appear that the Voting Rights Act requires this. Are there other regulations you are referring to?

[Zippyjuan]

If I was to be put in charge of districting, I would require that political affiliation be completely removed from where the lines are drawn and go with purely population and geography. Trying to balance the political makeup of a district- whether that is to favor one or another party or to try to force a balanced political district- is not natural and will contain bias either way. Draw the simplest lines which take into account geographical logic. Yes, you will have some districts which will tend to vote for one party or another. Yes, some will be more balanced. People of similiar income status tend to live closer together and people of similar incomes may also have a tendency to support one party or another. That does not mean a district was improperly drawn. If a person on one block can vote within a half mile and a person in the next block has to go five miles to vote, that is a poorly drawn districting. Make it convenient for the people living in it to go and vote. I think Driege says it simpler than I did, but you probably get the point.

[MozoVote]

Most people here would say districts "shouldn't be" set up to quarantine voters along racial dimensions. But N.C. has a history of getting into confrontations with the Justice Dept over these VRA districts. It's time consuming and exasperating to have a map thrown out by the courts, and continue to edit it and re-apply. Glen is just staring down the practical path, that the committee must go through.

I mostly agree with nate895. It's not necessary to re-invent the wheel here.

If you get on the committee, suggest to the chair that instead of an "independent commission" that several small independent teams be formed of 2 to 4 people each. They can get together at a computer, click maps together and brainstorm. Have each team return their ideas to the committee anonymously - and let the committee debate the submissions. Pick the two or three best ideas, and submit them back to the teams with feedback.

Let the teams reconsider the committe's comments, make fresh submissions, then the committe picks the best map.

I'm skeptical a computer program will pick up the nuances which are involved in making a successful plan. It's kind of like weather forecasting - computers can help model things, but people make the decisions.

[Original_Intent]

I disagree that balancing based on partisanship should have anything to do with it.

I do like the idea that they should be as compact as possible to avoid gerrymandering.

[Brian4Liberty]

Gunny, rather than trying to create balanced districts, which does have seem like a positive change from the status quo, I tend to think districts should be drawn without regard for the racial or political demographics (black vs white, Dem vs Rep). Rather, they should be drawn as the simplest polygons possible with equal numbers of constituents in each. What is your opinion of this?

I agree with your sentiment on objective Districts. At first I thought about simple polygons, but then revised that to what I call "Natural Boundaries". In other words, never unnecessarily break up towns, cities or counties. So a District would have to take up an entire city/town until it starts into a new city/town. Same with counties. So there would never be more than one fractional town, city or county in a given District.

So one District with County A and 1/2 County B would be OK. One District with 1/4 County A, 1/4 County B, 1/8 County C, and 3/8 County D would not be allowed. Might be easy to program that.

Also when a town/city/county is broken into two pieces, it would be on a major, existing boundary, like a river, freeway, or highway.

[nate895]

See my post above. It doesn't appear that the Voting Rights Act requires this. Are there other regulations you are referring to?

Like it or not, that is how the DoJ and SCOTUS interpret it.

[nate895]

I agree with your sentiment on objective Districts. At first I thought about simple polygons, but then revised that to what I call "Natural Boundaries". In other words, never unnecessarily break up towns, cities or counties. So a District would have to take up an entire city/town until it starts into a new city/town. Same with counties. So there would never be more than one fractional town, city or county in a given District.

So County A and 1/2 County B would be OK. 1/4 County A, 1/4 County B, 1/8 County C, and 3/8 County D would not be allowed. Might be easy to program that.

Also when a town/city/county is broken into two pieces, it would be on a major, existing boundary, like a river, freeway, or highway.

This is how most well-balanced states try to do it, and it works out decently. However, it is sometimes necessary to split up counties even further so that there are 2 1/2 counties, etc. This happens in cases where there are two bordering high-population density counties.

[MozoVote]

Yeah - North Carolina's constitution was written when the largest cities in the state were about 20,000 people, and there was not Federal ovsight of the districts. As a practical matter, even a fair map will probably break about a dozen counties.

And we have not even begun talking about the statehouse districts...

[Agorism]

I don't favor competitive districts because I don't like Democrats. Ron Paul district is so republican that it is nearly impossible for him to lose so long as he wins his primary.

I kind of like that.

[MozoVote]

Actually - Ron Paul's district was designed to be very difficult for a Republican to win. The Texas legislature and George W did not really want him around. Paul wins it by large margins because of his record of service to his constituents.

[Brian4Liberty]

So one District with County A and 1/2 County B would be OK. One District with 1/4 County A, 1/4 County B, 1/8 County C, and 3/8 County D would not be allowed. Might be easy to program that.


Clarified that a little. One County or City with 5 whole Districts inside it would be fine too, as long as no single District contains fractions of multiple cities or counties.

[nate895]

Clarified that a little. One County or City with 5 whole Districts inside it would be fine too, as long as no single District contains fractions of multiple cities or counties.

It still can be necessary for districts to have multiple county/city fragments. For example, in the Seattle Metro there are several high density population counties right next to each other: King, Snohomish, and Pierce. It is necessary for one district to be made of parts of Snohomish and King counties, one to be basically Seattle, another to be split between King and Pierce counties. There really is no other way to do it and maintain something close to mathematically ideal population.

[MozoVote]

Another reason counties get split up, is the Justice Dept very strongly takes the view that "One person one vote" trumps district lines in Federal elections. Districts need to balance very close, I think to within a half a percent of even.... or possibly better. They really don't care about whatever inconvenience it causes for political party caucuses.

There is a little more leeway in state elections but not a great deal. There have been court cases on that as well.

[GunnyFreedom]

I don't like the VRA requirements any more than anybody else. It isn't like nullification, if you ignore them, they WILL strike your map down in Federal court, and you have no recourse but to obey them. IIRC dist 12 was created after several Federal court rejects until they had a gerrymander that they liked.

WRT balancing towards zero partisanship, I don't like the idea of "safe" districts. At all. That just begs for corruption IMHO.

I do like the idea of "smallest polygons" but county borders are hard boundaries per the State Constitution. If it wasn't for the VRA then this process would be ridiculously simple -- group whole counties as close to same populations as possible and you are done.

My original idea before I learned all the requirements was community based districting, where primary communities get to have their own rep. That sounds more fair than anything to me.

That's why I am looking for an algorithm with neutralizing guidelines, otherwise there will always be bias.

WRT Mel Watt, I have learned that the FedGov will not allow him to be ungerrymandered (That one went, I believe, all the way to SCOTUS, who ruled that gerrymandering was necessary if it helps minorities, and they would reject any district that wasn't so gerrymandered), but we can compact the ^%$# out of his district so it diminishes his impact on the rest of the State.

Yeah, corruption corruption. :mad:

And SCOTUS is worse now than when that decision was handed down.

I'm not sure "smallest polygon" will work -- won;t that leave the last district as "all the rest of the space" ? maybe distributed area...

[driege]

Like it or not, that is how the DoJ and SCOTUS interpret it.

What I posted indicates that is not how SCOTUS interprets it.

[Matt Collins]

http://www.gifninja.com/Workspace/e0d38f1a-c546-4c6e-906d-73ea78dc369b/output.gif

[angelatc]

What I posted indicates that is not how SCOTUS interprets it.

I don't know about NC specifically, but there are several other instances that I'm aware of where the DOJ has mandated that certain states create minority districts. It is entirely possible that NC has one or two of those districts.

If I were drawing around that parameter, I'd be sure to focus on the Latino/Hispanic voters, because I think they're more likely to swing Republican.

[MozoVote]

N.C. has a Voting Rights Act 12th and 1st district. The 12th is the most egregious. Quoting Wikipedia:



Originally drawn in 1992 as a 64 percent black majority district stretching from Gastonia to Durham. It was very long and so thin at some points that it was no wider than a highway lane, as it followed Interstate 85 almost exactly, and was criticized as a gerrymandered district. The Wall Street Journal called the district "political pornography." The United States Supreme Court ruled in Shaw v. Reno, 509 U.S. 630 (1993) that a racial gerrymander may, in some circumstances, violate the Equal Protection Clause. Subsequently, the district was redrawn several times and was adjudicated in the Supreme Court on two further occasions. The current version has a small plurality of whites.

http://upload.wikimedia.org/wikipedia/en/9/95/NC12_109.gif

[itshappening]

Mel Watt (D-Bank of America) is an unbelievable example of gerrymandering

I do believe Gunny is right in wanting to make as many competitive districts as possible because the quality of the candidate and ideas and policies will prevail. Safe districts on either side are not good

However, I doubt Gunny will get a chance to shake things up as there is a lot of interest in the carve up and the current incumbents will make sure their fiefdoms are protected

[GunnyFreedom]

Mel Watt (D-Bank of America) is an unbelievable example of gerrymandering

I do believe Gunny is right in wanting to make as many competitive districts as possible because the quality of the candidate and ideas and policies will prevail. Safe districts on either side are not good

However, I doubt Gunny will get a chance to shake things up as there is a lot of interest in the carve up and the current incumbents will make sure their fiefdoms are protected

This all depends on whether I can maneuver my way onto the committee. :D

[GunnyFreedom]

Here is my 1st try by hand with 14 districts. I didn't even look at partisan leaning on this one, just tried to balance population alone:

http://away.144Ridgewood.net/filebin/house49rep/try1.gif

My chief problem with THIS map, ia that Alamance and Ashe counties have pretty much nothing in common. :(

[Rael]

I don't care if a court did say to gerrymander the district. Ignore them and draw the district fairly. If a court battle is needed, so be it. Make the other side at least have to fight.

[South Park Fan]

Does state law require that congressional districts be contiguous? If so, that ought to eliminate many possibilities right there.

[GunnyFreedom]

I don't care if a court did say to gerrymander the district. Ignore them and draw the district fairly. If a court battle is needed, so be it. Make the other side at least have to fight.

That'd be GREAT if it were up to me -- and I do fully intend to push that as my primary. Chances that such a plan would ever make it out of committee...... :(

So I also need a backup plan that WILL make it out of committee, or my purpose there will be merely entertainment for the other members.

[GunnyFreedom]

Does state law require that congressional districts be contiguous? If so, that ought to eliminate many possibilities right there.

Yes. voting districts must be contiguous.

[MozoVote]

Nothing wrong with some thinking outside the box... but the south Mecklenburg delegation would not want to be included in Mel Watt's district. Ruth Samuleson and Ric Killian will not vote for that.

Greensboro and Winston Salem usually have separate representatives due to historical competition between the cities.

I may try my own hand at a "fair" map when I get home tonight. I already made a "Democrat Nightmare" map yesterday that would only leave the 12th, 8th, and 1st as winnable. LOL! Gerrmandering is more fun! :D

[GunnyFreedom]

Nothing wrong with some thinking outside the box... but the south Mecklenburg delegation would not want to be included in Mel Watt's district. Ruth Samuleson and Ric Killian will not vote for that.

Greensboro and Winston Salem usually have separate representatives due to historical competition between the cities.

I may try my own hand at a "fair" map when I get home tonight. I already made a "Democrat Nightmare" map yesterday that would only leave the 12th, 8th, and 1st as winnable. LOL! Gerrmandering is more fun! :D

Yeah, that knowledge of historic issues certainly helps, and Ric Killian is certainly going to be one of our biggest allies in the House, even if I am uncomfortable considering a specific person when redistricting. Bottom line is if a Rep is really hurt by redistricting it will never pass the vote.

Well, as long as the redistricting is being done by HAND instead of automated that is :D

My goal here was to maintain the VRA districts in the smallest polygon possible -- ie encircling just the deeply urban downtown areas. That way, as few people as possible will get stuck in a VRA district who did not want to be.

So one of my primary considerations is that while the FedGov will require 1 or 2 VRA districts, I would prefer that most of the people IN the VRA district are the people who actually WANT to be in that district. I know right? Trying to see to it that the people on the ground are represented by the representative they actually want. Novel concept huh?

Anyway, this is going to take LOTS of work... :(

I will be letting Thom and Dale and Skip know that I really want on the Redistricting Committee today.

[MozoVote]

They really should put you on the committee. I'm not too hopeful, redist will be something a lot of them will want a piece of... But as a GOP rep in a difficult district, they need some people on the committee that are prepared to make honorable suggestions that won't embarras the party.

If the redist outcome becomes too extreme, the Dems could do in 4 or 6 years what the Texas GOP did. There is *NO* limitation to preventing redistricting in between the censuses.

If you can't get on redist, try for rules. These are committees where you can demonstrate a willingness to be sensible and non-political, and really act in the interest of the public.

[MozoVote]

OK, so I picked up some voter registration stats and a few things jump out to me right away. (I probably should use population figures, but registered voters is a good proxy among large samples, such as by Congressioanl District.)

http://www.jwpcivitasinstitute.org/voter-registration-statistics-district-u-s-congress

Ideally, each district would be 1/13 of the population. Districts 4 and 9 are each over 20% beyond that proportion, so they are the most apt to experience a major adjustment. The other districts are hovering reasonably near around 1/13, except for district 1 which has shrunk to be negative 14% underneath the benchmark and district 8 which is negative 9.6%.

Although there has been speculation over the last few years that NC could get a 14th district, it seems like that talk has died down recently. It's plausible, but no longer expected. (NC just barely got it's 13th in Y2000, so despite in-migration, we have some work to do in earning the 14th.)

[MozoVote]

I took two whacks at this - am still not satisfied, but it's getting better. Balancing the factors is harder than you might think.

IE, create districts that are more competitive, but not so much so that they'd agitate the existing congresspeople, or cause a GOP dominated legislature to balk... include two VRA districts, keep the others compact ... yet avoid splitting too many counties into 3 districts, and preferably not split any cities (badly) ... and also keep in mind commuting patterns, geography and historical context.

[GunnyFreedom]

I took two whacks at this - am still not satisfied, but it's getting better. Balancing the factors is harder than you might think.

IE, create districts that are more competitive, but not so much so that they'd agitate the existing congresspeople, or cause a GOP dominated legislature to balk... include two VRA districts, keep the others compact ... yet avoid splitting too many counties into 3 districts, and preferably not split any cities (badly) ... and also keep in mind commuting patterns, geography and historical context.

I have been working this problem a LOT, and it is tougher than ppl can imagine trying to do it by hand. I appreciate your help more than I can say, as you certainly have a lot more background data pertinent to the task than I do. Thanks!

[MozoVote]

I'll go ahead and post what I have... Sleep on it and revisit tomorrow. It's probably not worth obsessing over anyway without 2010 census data.

I'm not satisfied with it partially because it makes district 2 slightly Obama leaning (jeapordizing Ellmers, who just finished a tough race.) It also slightly flips district 6 (Coble, who is personally secure, but making this district Dem leaning would never pass committee - for fear of what happens when he retires.)

The districts do become more contiguous though, and I boxed Mel Watt's district down to 4 counties.

http://sdarwag.home.mindspring.com/redist/Nov22_cong_dist.jpg

[GunnyFreedom]

I'll go ahead and post what I have... Sleep on it and revisit tomorrow. It's probably not worth obsessing over anyway without 2010 census data.

I'm not satisfied with it partially because it makes district 2 slightly Obama leaning (jeapordizing Ellmers, who just finished a tough race.) It also slightly flips district 6 (Coble, who is personally secure, but making this district Dem leaning would never pass committee - for fear of what happens when he retires.)

The districts do become more contiguous though, and I boxed Mel Watt's district down to 4 counties.

http://sdarwag.home.mindspring.com/redist/Nov22_cong_dist.jpg

That's hot! It's better than what I have so far, except that mine tried a lot harder to preserve whole counties -- but then that makes it almost impossible to balance population and partisanship. I like yours better. I also like the fact that yours (completely unintentional I am sure) actually lets me vote for BJ Lawson. :D

[danda]

Mozo -- what did you use to generate that map? By hand or programmatically?

Anyway, I do have a tecnical suggestion regarding the OP.

The PostGIS extension to PostgreSQL DB supports geographic data calculations.

Here is an article from IBM explaining how to import city and county data and then make a mashup viewable in google earth.

http://www.ibm.com/developerworks/opensource/library/os-kmlservice/index.html?ca=drs-

Here is another basic intro to importing county data:

http://jduck.net/2009/01/30/getting-to-know-postgis-part-ii/

QGis is another viewer:

http://www.qgis.org

AFAIK, all of the above are free / open source.

Maybe you guys are already way ahead. If not, perhaps it can be a starting point.

[MozoVote]

I am using "Daves' Redistricting" as is Gunny.

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

This does become a bit easier if I let Watt's district enclose High Point... but then that's 5 counties of his "snake of a district..." and I want to make a clearer case that his district does NOT need to be such a long, thin ribbon. :(

I could also try running the 1st district, which is also VRA, into downtown Raleigh. That would allow for more McCain votes in the 2nd and 4th and they would be more competitive. Because the 1st has either lost population or grown very slowly, there actually is a non-partisan argument that it should not be penalized for being so rural and that it should have one of the fast growing urban areas in it.

Maybe a "Plan A" and "Plan B" would be a good idea-- don't get too wedded to one plan, just come up with better solutions than currently exists, to allow for some a la carte thinking.

[MozoVote]

I put some fine tuning into this tonight, and I'm beginning to think this plan is working out fairly well. If only we'd been doing the districts like this in the first place!! :rolleyes: All the districts do balance within 4000 persons in Dave's App.

Not much change to 1 (safe Dem) or 5 and 10 and 11 (safe R's).

Polished up district 2 to be 50% McCain (was 47% in the election) so that gives newcomer Marie Ellmers (R) some safety.

District 3 drops to 54% McCain from 61%, but it's currently held by long time incumbent (R) Jones so unlikely to flip. GOP leaning, but arguably competitive.

District 4 wraps around Wake county better to express a "Raleigh commuter" demographic, and I stopped splitting Johnston county with it. 48% McCain versus 36% in the actual election. It's much more competitive for David Price (D) to try and hold.

I trimmed district 6 some more. Now a 50/50 district, which could cause some grumbling - but this district has strong historical ties dating back 50 years or so, centering on Guilford and Randolph counties. Now High Point and Greensboro have contiguous representation, without Mel Watt's district messing it up. I think it can be argued that demographic changes are simply being expressed within rational boundaries. Howard Coble will probably retire soon, so if a truly competitive district must emerge, this is a good place to put it.

District 7 held by McIntire (D) ... the McCain % drops to 47% from 52% so it's a safer district, but not deeply so, and more compact than what currently exists.

District 8 drops just a smidge to 46% McCain from 47%, held by moderate Larry Kissell, could still be competitive, although Dem leaning.

Distict 9 moves east of Charlotte. McCain votes drops a little to 57% but this is still pretty conservative ground for Sue Myrick (R) as she keeps all the south Charlotte and Union county voters that she's had before.

District 12 fits within 4 counties. 56% white, 44% nonwhite by the old census, but I think the 2010 census would show a closer balance. District 9 stays along side the whole east flank of it, so that each county can speak with Sue Myrick if they are having problems working with Mel Watt.

Dist 13 becomes a mildly (R) district at 51% McCain versus 40% in the election. Winston Salem gets completely represented, without Mel Watt's gerrmander protruding into it anymore. Brad Miller (D) may not be able to hold this district any longer.

http://sdarwag.home.mindspring.com/redist/Nov23_w.jpg

http://sdarwag.home.mindspring.com/redist/Nov23_e.jpg

[MozoVote]

Been thinking about this a bit more ... Although Ellmer's new district is geographically compact, it is true that it does not contain much land from her old "3 spoke" district.

I'm gonna try and wrestle up another competing version over the next few days. Maybe I should put Franklin and Johnston county back into her district and then re-work the others, and see what I come up with.

[GunnyFreedom]

That's amazing work! It's a really tough balance to strike - I know. I do not know where the current reps live exactly, so I imagine in some cases districts may have shifted out from under people which will be a tough sell. You know a lot more about where they live than I do though.

The new Dist4 has me able to vote for Lawson! Huzzah!

Man you rock, really you do.

Now the real PITA is going to be drawing legislative districts. What a nightmare!

I'm saving this map. Thanks again! I am sure that we will have access to a more powerful redistricting app in the State House. If I get on that committee, how do you feel about coming to Raleigh for a visit to help make maps on the official machine? :D ;)

[GunnyFreedom]

Been thinking about this a bit more ... Although Ellmer's new district is geographically compact, it is true that it does not contain much land from her old "3 spoke" district.

I'm gonna try and wrestle up another competing version over the next few days. Maybe I should put Franklin and Johnston county back into her district and then re-work the others, and see what I come up with.

The advantage to doing that, of course, is that whenever I toss my hat into the ring for US Congress I won't end up in a district already represented by BJ Lawson. :eek:

[MozoVote]

I'm sure once you're sworn in, you'll have no trouble getting all the data you want from the State Board of Elections! ;)

Honestly I think 2 or 3 competing plans is good. Then you can pitch them before the other Legislators and county commissions. Ask them to pick what they like. Take all that feedback home, and then lay out the "Bradley Plan" :cool:

[GunnyFreedom]

See: http://nc49.org/legis/?q=node/6

[Aratus]

OK, this is a very complicated problem. It's complicated enough to redraw the districts in a way that obeys State law while attempting to balance the partisanship towards zero. Add in the Federal Voting Rights Act, and it gets ridiculously complicated. The best idea I have at this point is to use a computer to crunch the millions of potential permutations until the most fair, legal, and balanced districts are created.

Here are the parameters:

1) The NC State Constitution says that counties cannot be split.

2) Federal Voting Rights Act pretty much requires at least 2 majority-minority districts, thus requiring that we violate #1, but do so as little as possible.

3) Voting Districts need to be as "compact" as possible, so no long snaking thin corridors connecting sections, or (given #2) as little as possible.

4) Population within districts needs to be as even as possible.

5) My personal preference is to make districts as competitive as possible, balancing partisanship towards zero. That will be impossible in the two districts required by #2 above.

NC will certainly have 13 congressional districts an possibly 14, so maps need to be produced for both eventualities.

NC will also be redrawing the 120 State House Districts and the 60 State Senate Districts, so maps will need to be produced for legislative redistricting as well.

MY THOUGHTS:

We could just use the results from the 2008 Presidential election to calculate partisanship, but to do so would clearly favor Republicans during redistricting. I would prefer to get precinct-by-precinct data for the last 5 elections and average them per precinct by % between Presidential, Gubernatorial, and Senate elections, and then average them per precinct across all 5 years. That would provide the most "fair" picture of precinct partisanship, and it should not be that difficult to produce. I am sure I can call on the SBOE and get them to provide that data to me as a State House Rep. A couple hours worth of spreadsheet work could produce the detailed 5 year partisanship data, once we have the base data to work with. Apply the resulting % in each precinct to precinct population, and you have the person-by-person data necessary for balancing the districts.

Once partisan data is plugged into the precincts, the problem gets complicated and will require the computer. The program in question will need to group counties by proximity, precincts by county, as well as by precincts by proximity.

My guess is that the two VRA majority-minority districts will need to be drawn by hand -- too many variables to program easily once you consider that almost every aspect of those two districts will be built on "exceptions" from the primary algorithms. If we were to try and get the computer to map those as well, all VRA district precincts need to share linear borders, be contiguous, and less than 50% white.

All districts will need to be composed of counties of precincts that share linear borders and are contiguous.

Myself and another member of RPF's are currently working to produce fair and legal models by hand, but the balancing act is incredibly tough.

So I am calling for help from a major mondo super computer programmer to help set up a program that can analyze the data and produce a map of voting districts that include the 2 VRA districts, and then balance the remaining precincts as evenly a possible by population, and partisanship, maintaining contiguous counties as much as possible.

I can see possible ways to do it, but writing the program is way beyong my capabilities.

Once we have the partisanship data by precinct, and we remove the two VRA precincts from the mix, the remaining precincts will yield a partisanship of X%/Y% (say 48%/52%). Every single remaining district then should end up as close to 48/52 as mathematically possible given the parameters mentioned above. That way, partisanship is dispersed as evenly as conceivably possible.

So please help -- hard core programmers!

Do not underestimate the critical importance of this effort. Your help will help BJ Lawson win. Your help will help Glen Bradley win when he decides to run for Congress. This effort could directly put two more Ron Paul's into Congress.

If you are unable to do this, but know someone who can, PLEASE we really need your help!

thanks!


GUNNY! its a thought! bright idea! when you sit in your local legislature,
are you going to create the new district you could possibly run in???
up here in the baystate, we just had our HOUSE delegation go from
ten Democrats to possibly nine by 2013 becuz N.E is not growing fast!

[Aratus]

if the new district thusly carved out is value neutral, clearly at times by a
musical chairs logic this impells a very confident veteran politician by
dsplacement theory to move over into something a political machine
can rise up out of! i am now thinking through the very idea that our
governor gerry gave his name to all serpentine political districts!!!

beacon hill now has the ball in their own court as they all quietly play
out the game of favorites. do they all hope someone resigns and runs for
the senate against sen. brown or do they divide up a Democratic stronghold
in order to hold off our local TEA PARTY? we have the opposite scenario up here!