Ohio Redistricting Contest

[dpmapper]

Districts 94, 95, and 99 form Senate District 5. If necessary I will add the 1 precinct to connect that touch point. District 99 is 80% Democratic while 94 and 95 are 60% Republican. Tossup Senate district with a Democratic (black) incumbent.

District 96, 97, 98 form Senate District 6. District 98 is a tossup while 96 and 97 are 55 and 60% Republican. Safe R Senate district.

80% D plus two 60% R's is a lean D district, 53-47ish.  Also, are you sure there is a D incumbent?  The 5th district elected an R in 2010.  http://en.wikipedia.org/wiki/Bill_Beagle

[jimrtex]

Approach OH like you would MI. Both states have detailed requirements for building districts. The requirements are different and OH only has their requirements apply to the legislature, however.

What were your thoughts on breaking ward boundaries to create majority black districts? I started in Hamilton County, which is much easier than the northeast, and you can create a 4-2-1 delegation there by connecting the white areas of eastern Cincinnati with the white areas of western Cincinnati.

Butler County is of course very easy, and Warren + excess in Butler creates 2 house districts. Linking 2 in Warren + the 7th in Hamilton should be an option to create a Republican Senate district.

My thought is that if it may be the case that a majority-minority district would be required, then it takes precedence over political boundary integrity. So, in Cleveland, where I can follow ward boundaries without sacrificing minority districts I do so. In Hamilton, I can barely create a senate district with over 50%. That forces me to ignore ward boundaries in Cinci.

Under the Ohio Constitution, splitting of wards is only permitted if it is non-feasible to have equal populated districts otherwise, and the only reason you can use whole wards is because it is not feasible to avoid splitting large cities.

Presumably, city wards represent local communities of interest.  Can they be cracked apart and joined with disparate communities simply in the interest of some arbitrary racial target.  Recalled that in LULAC v Perry, the Supreme Court did not count a district as a VRA district, even though it had a majority Hispanic population.

[muon2]

Approach OH like you would MI. Both states have detailed requirements for building districts. The requirements are different and OH only has their requirements apply to the legislature, however.

What were your thoughts on breaking ward boundaries to create majority black districts? I started in Hamilton County, which is much easier than the northeast, and you can create a 4-2-1 delegation there by connecting the white areas of eastern Cincinnati with the white areas of western Cincinnati.

Butler County is of course very easy, and Warren + excess in Butler creates 2 house districts. Linking 2 in Warren + the 7th in Hamilton should be an option to create a Republican Senate district.

My thought is that if it may be the case that a majority-minority district would be required, then it takes precedence over political boundary integrity. So, in Cleveland, where I can follow ward boundaries without sacrificing minority districts I do so. In Hamilton, I can barely create a senate district with over 50%. That forces me to ignore ward boundaries in Cinci.

Under the Ohio Constitution, splitting of wards is only permitted if it is non-feasible to have equal populated districts otherwise, and the only reason you can use whole wards is because it is not feasible to avoid splitting large cities.

Presumably, city wards represent local communities of interest.  Can they be cracked apart and joined with disparate communities simply in the interest of some arbitrary racial target.  Recalled that in LULAC v Perry, the Supreme Court did not count a district as a VRA district, even though it had a majority Hispanic population.

The key fact in LULAC was the linking of two Latino areas separated by 300 miles and that shared no other recognizable bond other than ethnicity. On top of that, there was evidence that a more tightly knit Latino district could have been created. IL-4 in the 1990 and 2000 remaps is a counter-example where the courts specifically allowed the linking of disparate communities by a thread because no other means could provide for a district where Latinos could elect a candidate of choice.

In any case it would be white-majority areas that would be linked by a ward-splitting strip in Hamilton. OH already does this in Franklin county to link Bexley to Whitehall.

[jimrtex]

I think we have been going at this from the wrong angle.  It is not necessary to draw unconstitutional house districts in NE Ohio.  

It is necessary to group them among senate seats in an unconstitutional manner.  And just because Cuyahoga and Lake counties have a whole number of house districts, doesn't mean that it is their house seats that are not grouped in unconstitutional senate seats.

So starting with Trumbull we draw the House seats:

H1. Trumbull (1017)
H2. Trumbull (788)-Ashtabula(229)
H3. Ashtabula(642)-Geauga(375)
H4. Geauga(426)-Portage(368)-Stark(223)
H5. Portage(1017)

I assume we can get across Portage N-S along the eastern edge without splitting more than one town (42,000 people).  Checked it.  The two eastern tiers of townships is about 31,000.  We can get the portion of Stark from Alliance.

We have 11 districts in Cuyahoga, 2 in Lake, 4 in Summit, 1 in Medina, and one Summit-Medina.

Senate districts:

(A) Split Lake and Trumbull

1-4.  Cuyahoga 11, Lake 1.
5.  Lake 1, H2, H3
6. H1, H4, H5
7-8. Summit 4, Medina 1, Summit-Medina

(B) Split Lake

1-4. Cuyahoga 11, Lake 1
5. Lake 1, H1, H2
6. H3, H4, H5
7-8. Summit 4, Medina 1, Summit-Medina

This requires that H2 cross Ashtabula from Trumbull to Lake, along the Pennsylvania border and then along the shoreline, and only split one town.  I doubt it is possible.  Not even close, even if we skipped the city of Ashtabula.  So (B) is out.

(C) Double split Cuyahoga and Summit

1-4.  Cuyahoga 10, Lake 2.
5. Cuyahoga 1, Medina 1, Medina-Summit
6. Summit 3
7. Summit 1, H4, H5
8. H1, H2, H3

So S7 is Portage, Summit(part), Stark(part), Geauga(part)
And S8 is Trumbull, Ashtabula, Geauga (part)

This has the advantage of respecting the ratio of representation for whole senate districts in Cuyahoga and Summit counties (11.06).

(D) Split Cuyahoga and Lorain

1-4.  Cuyahoga 10, Lake 2.
5. Medina 1, Medina-Summit, Summit 1
6. Summit 3
7. H4, H5, and Jefferson-Carroll-Harrison
8. H1, H2, H3
9. Lorain 2, Cuyahoga 1
10. Lorain-Huron, Ottawa-Erie, Sandusky-Seneca

S7 from Steubenville to near Cleveland is a bit extreme, and we might not be able to get across Stark.  So (D) is out.

I kind of like (C) more than (A).   Cuyahoga and Summit still have whole senate seats, and the other counties are not split (except Geauga).

[muon2]

I appreciate the idea of violating the senate rules for counties as opposed to the house rules. It's an interesting approach. I'm not sure that one has precedence over the other.

Does plan (C) still provide two majority BVAP senate districts?

[krazen1211]

Districts 94, 95, and 99 form Senate District 5. If necessary I will add the 1 precinct to connect that touch point. District 99 is 80% Democratic while 94 and 95 are 60% Republican. Tossup Senate district with a Democratic (black) incumbent.

District 96, 97, 98 form Senate District 6. District 98 is a tossup while 96 and 97 are 55 and 60% Republican. Safe R Senate district.

80% D plus two 60% R's is a lean D district, 53-47ish.  Also, are you sure there is a D incumbent?  The 5th district elected an R in 2010.  http://en.wikipedia.org/wiki/Bill_Beagle

Hmm, I suppose you are right.

Either way, Greene County is growing faster than Dayton, and more importantly, the Democratic primary would be dominated by urban blacks. It's a decent shot to hold the seat.

Looking back, I wish that Montgomery could be linked to Warren County, but it fits so nicely with Greene.

[krazen1211]

Hamilton County and surroundings.




Districts 8, 9, and 10 form Senate district 7. Safe R.

Districts 1, 2, and 6 form Senate district 8. Safe R.

Districts 3, 4, and 5 form Senate district 9. District 3 is a tossup wholly in Cincinnati/Norwood. District 4 and 5 are safe Democratic majority black districts. Recombining these by wards will simply yield district 3. The Senate seat is safe D.

[dpmapper]

Hmm, I suppose you are right.

Either way, Greene County is growing faster than Dayton, and more importantly, the Democratic primary would be dominated by urban blacks. It's a decent shot to hold the seat.

Looking back, I wish that Montgomery could be linked to Warren County, but it fits so nicely with Greene.

Beagle is from Miami County, so in an effort to keep him in that seat, I'd put the Montgomery leftovers with Preble and then make a Darke-Miami house seat (splitting one of them).  Added benefit is that these areas are even redder than Greene County is, and this puts the western/northwestern counties to some good use other than simply vote-sinking them.  I'll work on a map doing that. 

What are the stats on your Hamilton house seats, particularly 2 and 6? 

[krazen1211]

Hmm, I suppose you are right.

Either way, Greene County is growing faster than Dayton, and more importantly, the Democratic primary would be dominated by urban blacks. It's a decent shot to hold the seat.

Looking back, I wish that Montgomery could be linked to Warren County, but it fits so nicely with Greene.

Beagle is from Miami County, so in an effort to keep him in that seat, I'd put the Montgomery leftovers with Preble and then make a Darke-Miami house seat (splitting one of them).  Added benefit is that these areas are even redder than Greene County is, and this puts the western/northwestern counties to some good use other than simply vote-sinking them.  I'll work on a map doing that. 

What are the stats on your Hamilton house seats, particularly 2 and 6? 

District 1: 71% R
District 2: 56% R
District 3: 52% D
District 4: 85% D
District 5: 78% D
District 6: 56% R
District 7: 59% R

I chose Greene because the numbers worked, but good point. Then you can use Greene County to neutralize some of the rurals closer to the Ohio river.

I tend to trust rurals less than I do suburban Republicans. And Greene is growing faster than Preble/Darke. That said your method probably works better.

[jimrtex]

I appreciate the idea of violating the senate rules for counties as opposed to the house rules. It's an interesting approach. I'm not sure that one has precedence over the other.

Does plan (C) still provide two majority BVAP senate districts?

I don't think so.  You need a white district to go with Lake, which means you need to come around the tip of southeast Cleveland.  So you would end up with the same as my current plan, but simply regrouping the western suburbs.

[dpmapper]

Drafts of Columbus, Dayton, and a tweak of Cincinnati:

First, Columbus:



Similar to krazen's plan, numbers are slightly more favorable: 

The puke green finger is 37% Republican, and will be attached to two outer county districts that krazen had at 64 and 60 GOP.  Still a likely GOP senate seat. 

Bright green in the NE: 54% Republican
Blue E/SE: 46.4% Republican
Purple south: 51.4% Republican. 
These three together are a very slight lean GOP senate seat. 

Orange west: 45% Republican
Blue west: 53% Republican
Magenta NW: 53.8% Republican. 
Again, a very slight lean GOP senate seat. 

So the GOP goes from having 3 house seats in Franklin to potentially 4, and has probably improved shots at keeping their senate seats, at the cost of weakening a senate seat in the collar counties. 

Cincinnati:



The yellow district is the same as before: 56.6% Republican, should be enough to take out the Dem incumbent. 

I noticed that the current map divides Cincy proper up into 5 districts (in fact, one GOP incumbent lives in Cincy) so there's no need to be too delicate, apparently.  My map only has 4, so it's an improvement.  One can unpack the western suburbs enough to flip one more house seat: the pink river district is now at 56.3% Republican.  The teal, grey, and cyan are still overly safe at 61.1, 59.8, 61.9 respectively.  The two remaining Dem districts are majority black.  Goes from 4-3.5 Dem to 5-2 GOP in the House delegation (senate stays 1-1 + a GOP leftover), and doesn't even need to use part of blood-red Warren County. 

Dayton:


Here's one way to do it. 

Yellow is 52.3% Republican
Pink is 50.1% Republican
Green is 60.9% Republican
Together this should be a safe senate seat.

Orange is a majority black, 20.9% Republican seat.
Bronze is a 60.8% Republican seat
and green up top (Darke, Mercer, and part of Miami) is 66.4% Republican.  Together this is a tossup senate seat. 

Currently in Montgomery there are 3 GOP seats to 2 Dem seats.  This endangers one or two of the GOP house seats (eviscerating one Dem seat) in exchange for making the second senate seat (Beagle's: the one that takes in inner city Dayton + the outer counties) quite a bit safer (but not safe) for the GOP incumbent. 

If the Ohio GOP decides that protecting house seats takes priority, then I'd unpack the bronze district a little bit.  Beagle still gets a safer district than his current one but it will be lean D strictly by the numbers.  One D house district still goes poof, and the other seats will be lean R at least. 

[jimrtex]

Under the Ohio Constitution, splitting of wards is only permitted if it is non-feasible to have equal populated districts otherwise, and the only reason you can use whole wards is because it is not feasible to avoid splitting large cities.

Presumably, city wards represent local communities of interest.  Can they be cracked apart and joined with disparate communities simply in the interest of some arbitrary racial target.  Recalled that in LULAC v Perry, the Supreme Court did not count a district as a VRA district, even though it had a majority Hispanic population.

The key fact in LULAC was the linking of two Latino areas separated by 300 miles and that shared no other recognizable bond other than ethnicity. On top of that, there was evidence that a more tightly knit Latino district could have been created. IL-4 in the 1990 and 2000 remaps is a counter-example where the courts specifically allowed the linking of disparate communities by a thread because no other means could provide for a district where Latinos could elect a candidate of choice.

In any case it would be white-majority areas that would be linked by a ward-splitting strip in Hamilton. OH already does this in Franklin county to link Bexley to Whitehall.

That is an interesting connector.  Columbus wards are relatively small, because they aren't used for electing the city council.  It also looks like wards have been drawn in a way that makes it feasible to link the fragmented townships.  You can include a Columbus ward or two with one of the townships and gather all the township in one district.

Back to the connector.  There is a Columbus ward with 6 precincts between Bexley and Whitehall.  They're in a 2x3 grid, 2 east-west and 3 north-south.  The connector is made up of the northern 2 precincts, and one of the central ones, though it isn't needed for the connector.  The 3 connector precincts are majority BVAP, while those not in the connector are much less (38%, 18%, 12%).  Maybe it wasn't so much to link white areas, but to build up the black population in a 3rd house district.  Whitehall in particular is not a white enclave.

Did you say that you had a majority BVAP in Franklin County?  Or was I confusing that with Hamilton?  Using precincts, if you take those that have the highest BVAP% and represent 3/10 of the total population, the composite BVAP is 50.3%.  So I think you would end a bit short when you start using towns and wards.

Franklin is less segregated than Cuyahoga, with the the threshold percentage around 20%, so I suspect that you can get in the mid to high 40s, without being extremely restricted on which areas you choose.

On the other hand, if you take the precincts representing 2/10 of the population (2 house seats), the BVAP% is 61.1%, which suggests that two majority BVAP districts can be drawn without too much difficulty.

I'm inclined to drop the existing 3 Franklin districts that are in the 95% to 105% range.  They form a peninsula coming from the Delaware line down through central Columbus.  Keeping them would mean that you would have a ring of towns running from Westerville clockwise to Dublin (12:30 to 10:30) that would have to be sliced into 7 districts with no flexibility in where boundaries have to be drawn.  If you remove the peninsula, you have more flexibility to push into Columbus to get to the right population totals.

The census VTDs appear to match the current precincts in Franklin county, sorta.   Back in 1990, the census bureau defined census blocks for the entire country, based on visible roads etc.  When city limits and the like crossed a census block,  the census bureau defined tabulation blocks, which were subdivisions of census blocks, and were used for tabulation city populations etc.  This was kind of messy, since the census blocks were intended to be the atomic area.

For 2000, the census bureau redefined census blocks to be defined by visible features and political boundaries.  So a 1990 census block divided by a city limit, became two 2000 census blocks.  But in Franklin County at least, it looks like the VTDs were not redefined.  The city and town limits follow block boundaries, even though where they are not coincident with visible features like streets.  But the VTDs appear to continue to use the 1990 census blocks.  So the VTDs for a town are not coincident with a town, even though the voting precincts are.

ps If you are in pan mode, and you hold the shift while doing a left-click and drag, you have a marquee zoom.

pps why are the blocks in Cleveland so long?

[jimrtex]

I think I've figured out the layout for Franklin County, a senate district with a 44.0% BVAP, and two house districts with a majority BVAP.  The districts are much more compact than the current districts, and respect towns more.

It probably is not possible to avoid dividing two townships between districts if Columbus is included.

[muon2]

I think I've figured out the layout for Franklin County, a senate district with a 44.0% BVAP, and two house districts with a majority BVAP.  The districts are much more compact than the current districts, and respect towns more.

It probably is not possible to avoid dividing two townships between districts if Columbus is included.

That's what I found. I created a black-majority SD within Franklin. However, some pairs of the 10 HDs in Franklin had to share townships in addition to sharing Columbus city. I found that the split townships included no splits of contiguous portions of townships. Even so, I assume this is a violation of the constitution, but it is perhaps justified to further the needs of the VRA. Someone would have to show that dividing an already disconnected township is more important than the VRA to invalidate that sort of split.

In Hamilton, it's not hard to get two majority-black HDs without splits of anything except Cincinnati. However, I had to split a fully connected township to get a black-majority SD in the county. A black-majority HD in Summit also requires splits of connected municipalities beyond Akron.

[jimrtex]

I think I've figured out the layout for Franklin County, a senate district with a 44.0% BVAP, and two house districts with a majority BVAP.  The districts are much more compact than the current districts, and respect towns more.

It probably is not possible to avoid dividing two townships between districts if Columbus is included.

That's what I found. I created a black-majority SD within Franklin. However, some pairs of the 10 HDs in Franklin had to share townships in addition to sharing Columbus city. I found that the split townships included no splits of contiguous portions of townships. Even so, I assume this is a violation of the constitution, but it is perhaps justified to further the needs of the VRA. Someone would have to show that dividing an already disconnected township is more important than the VRA to invalidate that sort of split.

In Hamilton, it's not hard to get two majority-black HDs without splits of anything except Cincinnati. However, I had to split a fully connected township to get a black-majority SD in the county. A black-majority HD in Summit also requires splits of connected municipalities beyond Akron.

This is my Franklin plan.  The discontinuities have been fixed.

https://districtbuilder.drawthelineohio.org/districtmapping/plan/827/view/

I'm rather pleased.  The statistics weren't being updated, and District 26 was the last remnant.  I started out assigning Franklin to one district, and the donut around Franklin in another.  I then started drawing each district.  So when I was drawing 25, what was left was 26.  Suddenly 25 went from blue to white, but 26 went from brown to blue.

I had to log off and log back on again to get statistics.  26 is minus 2.7%.

No Columbus wards are split.

There are three intentional township splits, and two of them are based on city boundaries.

Miflin: Gahanna is separated from the remainder.

Jefferson: Reynoldsburg is separated from the remainder, but unified with the main portion of the city in Truro.

Clinton: The fragment west of OSU is in one district while the part in the north central is in another.  They must be in opposite corners of the original township.

Other incidental fragments have just a few hundred persons total.

I don't think the compactness is working.  20 gets a brown tint if you specify compactness as the district theme.  That might be possible if they are using a perimeter measurement, since it uses the city limit of Upper Arlington and the main N/S throughfare.

But if you look at the compactness scores it doesn't do so good.  They may be using an Iowa scort (EW to NS ratio).  How else to understand the high score for 17?

The senate district is 45.7% BVAP.

[muon2]

I think I've figured out the layout for Franklin County, a senate district with a 44.0% BVAP, and two house districts with a majority BVAP.  The districts are much more compact than the current districts, and respect towns more.

It probably is not possible to avoid dividing two townships between districts if Columbus is included.

That's what I found. I created a black-majority SD within Franklin. However, some pairs of the 10 HDs in Franklin had to share townships in addition to sharing Columbus city. I found that the split townships included no splits of contiguous portions of townships. Even so, I assume this is a violation of the constitution, but it is perhaps justified to further the needs of the VRA. Someone would have to show that dividing an already disconnected township is more important than the VRA to invalidate that sort of split.

In Hamilton, it's not hard to get two majority-black HDs without splits of anything except Cincinnati. However, I had to split a fully connected township to get a black-majority SD in the county. A black-majority HD in Summit also requires splits of connected municipalities beyond Akron.

This is my Franklin plan.  The discontinuities have been fixed.

https://districtbuilder.drawthelineohio.org/districtmapping/plan/827/view/

I'm rather pleased.  The statistics weren't being updated, and District 26 was the last remnant.  I started out assigning Franklin to one district, and the donut around Franklin in another.  I then started drawing each district.  So when I was drawing 25, what was left was 26.  Suddenly 25 went from blue to white, but 26 went from brown to blue.

I had to log off and log back on again to get statistics.  26 is minus 2.7%.

No Columbus wards are split.

There are three intentional township splits, and two of them are based on city boundaries.

Miflin: Gahanna is separated from the remainder.

Jefferson: Reynoldsburg is separated from the remainder, but unified with the main portion of the city in Truro.

Clinton: The fragment west of OSU is in one district while the part in the north central is in another.  They must be in opposite corners of the original township.

Other incidental fragments have just a few hundred persons total.

I don't think the compactness is working.  20 gets a brown tint if you specify compactness as the district theme.  That might be possible if they are using a perimeter measurement, since it uses the city limit of Upper Arlington and the main N/S throughfare.

But if you look at the compactness scores it doesn't do so good.  They may be using an Iowa scort (EW to NS ratio).  How else to understand the high score for 17?

The senate district is 45.7% BVAP.

A plan with 11 house districts and 4 senate districts each over 50% BVAP is posted for sharing.

[jimrtex]

A plan with 11 house districts and 4 senate districts each over 50% BVAP is posted for sharing.

I think they are using two different compactness measurements.  If you set the district theme as compactness, you have 3 very non-compact districts, which indicates that a perimeter based measure is being used, because they are districts with tentacles.

The scores may be using the correct measures for the contest.  Notice that the Dayton district scores relatively high.   And Wayne gets pretty close to the 2/π for a square with circle circumscribed.

I think the Akron seat is dubious.  Just because you have to follow the tentacles of the city limits doesn't mean that you can stick a claw into the center of the city to grab some white voters.

Also you appear to have a systematic underpopulation bias.  8 of your 11 majority BVAP districts have a blue tint.  3 of the 28 other districts in those 6 counties are blue, and 3 are tan.  To the extent that you did not respect political boundaries and city wards in violation of the Ohio Constitution your map violates OMOV.

The contest rules should include a population equality measure, such as RMS deviation.

I am going to borrow your District 97.  My districts in the southern part of the state were the same as yours, but I had split the counties in the east a little bit differently, and had ended up splitting Holmes to add a bit to Ashland-Knox which had really bugged me.

I don't think you reduce the number of county cuts in the west by not keeping the existing districts.

I would switch a couple of townships to better balance the population in 78-82 and 92-93.  Once you've decided to split the county (Hardin and Washington) you might as well go for better equality rather than preserving a sense of COI with the county seats.

Have you been able to run the reports?  I have not been able to do so.

And how did you create your senate map?

I'll borrow the Butler-Warren-Hamilton senate seat.  I had blindly stuck the Clermont-Brown districts with Hamilton since 7+2 = 9, and perhaps because Montgomery-Greene was an easy 6, I was stacking the senate groups.

[muon2]

A plan with 11 house districts and 4 senate districts each over 50% BVAP is posted for sharing.

I think they are using two different compactness measurements.  If you set the district theme as compactness, you have 3 very non-compact districts, which indicates that a perimeter based measure is being used, because they are districts with tentacles.

The scores may be using the correct measures for the contest.  Notice that the Dayton district scores relatively high.   And Wayne gets pretty close to the 2/π for a square with circle circumscribed.

The numbers in the table seem to be Roeck in accordance with the competition. District Builder is designed for more than one state's contest and they may have set the highlighting tool for a another state and not changed it for OH. The evaluation page includes a length vs width compactness measure and that is not part of OH either.

You must be logged in to read this quote.

Akron was by far the toughest to draw. But, since this is a policy-based plan to highlight the limits of rough proportionality under the VRA, it seemed appropriate to push the limits. There's a lengthy explanation file for the plan but it is too long for the plan notes. It would be nice if the file could be part of the shared plan.

You must be logged in to read this quote.

A big part of that is due to the constitutional constraints. Hamilton county has enough population for only 6.88 districts but that can be divided into seven districts with an average size 2% less than the ideal. Small districts are to be expected there. Anyway, both districts 31 and 32 have sufficient black population to draw from HDs 30 and 26 and still remain black-majority, so those are an easy fix if demanded. Similarly, in Cuyahoga district 12 could draw population from district 10 and go into the white tint.

You must be logged in to read this quote.

I agree that points should be deducted for districts that vary more than some predetermined amount. That would make OMOV a goal to be balanced with the others. Buut, that's not a factor here.

You must be logged in to read this quote.

It was very pleasing to get the SE OH counties grouped to require only one county split (92-93) among counties less than the size of a district. And no splits in the senate in SE OH.

You must be logged in to read this quote.

Is there a specific district that you think could be preserved in the map? I know that keeping 86 as is resulted in extra splits.

You must be logged in to read this quote.

Since the contest doesn't reward for population equality beyond 5%, I'd be more likely to focus on compactness and partisan composition.

You must be logged in to read this quote.

Total scores and reports have not been enabled in the software for OH. It's been a lower priority than fixing speed issues and geounit bugs.

You must be logged in to read this quote.

I asked and got a time-consuming script involving zip files and editing with Notepad. I've noted some issues with the script, and I hope that it is added to the web page for other users. PM me if you'd like my script.

You must be logged in to read this quote.

I actually start with regional groups of 4-6 SDs with whole counties only, then create the SDs. Those are divided and then iterated against the HDs to reach a final map.

[jimrtex]

Hamilton County preliminary version.

I started out trying to create an area in the central part of the county that was as high BVAP% as possible, and made up of whole townships and Cincinnati wards, equivalent to 3 house districts, with two districts to the west, and two to the east.  Getting rid of enclaves and balancing the two western and eastern parts gives a distribution about like I've drawn.

I then realized that the two western house districts couldn't form a senate district.  I played a little trying to cut across along the river, but that didn't seem to work out using whole wards.  So I decided to just try to create 3 BVAP plurality districts, with perhaps a majority in one.

The combination of Green and Coleraine avoids a split of townships, but it is rather odd that two townships each with 1/2 the population of a district can be combined.  That is why the further west district wraps around into southwestern Cincinnati.

Getting the populations to stay in range, in particular with the countywide shortage of 1.3% took some doing.  Other than Cincinnati, the only split townships are due to fragments.  Three enclaves of Springfield township in northern Cincinnati are split from the main part of the township and other fragments.  And a fragment of Columbia township on the eastern boundary with Clermont is also split off.

After looking at the map (the one I shared) I made some tweaks to HD 31, which brings it up to 54% BVAP, which would give me 7 majority BVAP districts (4 in Cuyahoga, 2 in Franklin, and 1 in Hamilton).  

If I could use the 3 central house districts to form a senate district, I am at 46%, but combining 30, 31, and 32 gives me 33%.   If I use 27, 29, and 31 I can do 35%, and an ugly district that stretches from the NW corner of the county to the east boundary.

I think it may be possible to cut across the northern edge of the county.  Taking Forest Park was not an obvious solution.  And it may be possible to cut across along the river after all.  I think those will end up in the low 40s.  I will try those.

[muon2]

Hamilton County preliminary version.

I started out trying to create an area in the central part of the county that was as high BVAP% as possible, and made up of whole townships and Cincinnati wards, equivalent to 3 house districts, with two districts to the west, and two to the east.  Getting rid of enclaves and balancing the two western and eastern parts gives a distribution about like I've drawn.

I then realized that the two western house districts couldn't form a senate district.  I played a little trying to cut across along the river, but that didn't seem to work out using whole wards.  So I decided to just try to create 3 BVAP plurality districts, with perhaps a majority in one.

The combination of Green and Coleraine avoids a split of townships, but it is rather odd that two townships each with 1/2 the population of a district can be combined.  That is why the further west district wraps around into southwestern Cincinnati.

Getting the populations to stay in range, in particular with the countywide shortage of 1.3% took some doing.  Other than Cincinnati, the only split townships are due to fragments.  Three enclaves of Springfield township in northern Cincinnati are split from the main part of the township and other fragments.  And a fragment of Columbia township on the eastern boundary with Clermont is also split off.

After looking at the map (the one I shared) I made some tweaks to HD 31, which brings it up to 54% BVAP, which would give me 7 majority BVAP districts (4 in Cuyahoga, 2 in Franklin, and 1 in Hamilton).  

If I could use the 3 central house districts to form a senate district, I am at 46%, but combining 30, 31, and 32 gives me 33%.   If I use 27, 29, and 31 I can do 35%, and an ugly district that stretches from the NW corner of the county to the east boundary.

I think it may be possible to cut across the northern edge of the county.  Taking Forest Park was not an obvious solution.  And it may be possible to cut across along the river after all.  I think those will end up in the low 40s.  I will try those.

Springfield township poses a problem, too. Though you could split the disconnected parts.

In any case your map shares my difficulty in that it becomes very uncompetitive when enhancing minority voting. OTOH, if you drop the SDs to 25-30% BVAP and have only one HD with over 40% BVAP, you can get 3 highly competitive HDs and two competitive SDs in Hamilton Co.

[jimrtex]

I think it may be possible to cut across the northern edge of the county.  Taking Forest Park was not an obvious solution.  And it may be possible to cut across along the river after all.  I think those will end up in the low 40s.  I will try those.

Springfield township poses a problem, too. Though you could split the disconnected parts.

In any case your map shares my difficulty in that it becomes very uncompetitive when enhancing minority voting. OTOH, if you drop the SDs to 25-30% BVAP and have only one HD with over 40% BVAP, you can get 3 highly competitive HDs and two competitive SDs in Hamilton Co.

North linked senate district.

It worked!.

The central senate district is 45.1% BVAP.

Deviation from the ideal population for Hamilton County:

1.3%, 1.3%, 0.9%, -0.5%, -0.5%, -0.5%, and -1.5%, overall 2.8%.

Other than the splits in Cincinnati, Springfield township is split, with a fragment north of Forest Park and a tiny fragment between Forest Park and Coleraine township in HD 27.  The main part of the township, and 4 smaller fragments are in HD 29.  Since HD 27 does not include any of Cincinnati, only one township is split between the districts.

Columbia township is split between HD-32 and HD-33.  Columbia widely dispersed fragments.  It is possible to include all within HD-33, but that would require including Mariemont in HD-33, which would in turn cut of Fairfax from HD-32.  Since HD-33 does not include any of Cincinnati, only Columbia is split between HD-32 and HD-33.

HD-30 is point contiguous based on census maps, but not if the Hamilton GIS web site is consulted.  The tip of Ward 19 is a couple of census blocks that are over a mile long running alongside Mill Creek and a railyard.  Because there are no streets across this area, the census blocks go further north than the ward actually does.  The whole tip could be cut off at its base on 8th Street and only shift about 200 voters.   Or one could remove an entire precinct and be back at the river (and remember Ohio precincts are tiny).

Alternatively, Ward 19, which includes the area along the river to the west could be swapped for Cheviot.  This would push HD-30 to around 3.0% negative deviation, but would reduce the split of Cincinnati to 4 districts.

Or one could instead notice that HD-31 looks like a moose; or that HD-29 has a 41.1% BVAP, and a 55.83% compactness score.

[muon2]

And how did you create your senate map?

The tips tab at the competition website has the script for building a senate plan from a house plan. It includes my suggestions, but I will say the process is slow.

[muon2]

I think it may be possible to cut across the northern edge of the county.  Taking Forest Park was not an obvious solution.  And it may be possible to cut across along the river after all.  I think those will end up in the low 40s.  I will try those.

Springfield township poses a problem, too. Though you could split the disconnected parts.

In any case your map shares my difficulty in that it becomes very uncompetitive when enhancing minority voting. OTOH, if you drop the SDs to 25-30% BVAP and have only one HD with over 40% BVAP, you can get 3 highly competitive HDs and two competitive SDs in Hamilton Co.

North linked senate district.

It worked!.

The central senate district is 45.1% BVAP.

Deviation from the ideal population for Hamilton County:

1.3%, 1.3%, 0.9%, -0.5%, -0.5%, -0.5%, and -1.5%, overall 2.8%.

Other than the splits in Cincinnati, Springfield township is split, with a fragment north of Forest Park and a tiny fragment between Forest Park and Coleraine township in HD 27.  The main part of the township, and 4 smaller fragments are in HD 29.  Since HD 27 does not include any of Cincinnati, only one township is split between the districts.

Columbia township is split between HD-32 and HD-33.  Columbia widely dispersed fragments.  It is possible to include all within HD-33, but that would require including Mariemont in HD-33, which would in turn cut of Fairfax from HD-32.  Since HD-33 does not include any of Cincinnati, only Columbia is split between HD-32 and HD-33.

HD-30 is point contiguous based on census maps, but not if the Hamilton GIS web site is consulted.  The tip of Ward 19 is a couple of census blocks that are over a mile long running alongside Mill Creek and a railyard.  Because there are no streets across this area, the census blocks go further north than the ward actually does.  The whole tip could be cut off at its base on 8th Street and only shift about 200 voters.   Or one could remove an entire precinct and be back at the river (and remember Ohio precincts are tiny).

Alternatively, Ward 19, which includes the area along the river to the west could be swapped for Cheviot.  This would push HD-30 to around 3.0% negative deviation, but would reduce the split of Cincinnati to 4 districts.

Or one could instead notice that HD-31 looks like a moose; or that HD-29 has a 41.1% BVAP, and a 55.83% compactness score.

I like it, but it won't score well for competitiveness.

[jimrtex]

And how did you create your senate map?

The tips tab at the competition website has the script for building a senate plan from a house plan. It includes my suggestions, but I will say the process is slow.

I got an e-mail on how to renumber the CSV file.

I had looked at the CSV file in a spreadsheet, but it was only for Cuyahoga and the 5-single county districts, and had 45,000 blocks.  I don't know if my spreadsheet can handle 500,000 rows and not mess up the formatting of a text file.

[jimrtex]

My senate and house plans

https://districtbuilder.drawthelineohio.org/districtmapping/plan/999/view/

https://districtbuilder.drawthelineohio.org/districtmapping/plan/1000/view/