Local vs regional road connections

[muon2]

On the Buncombe county GIS web site, the ERSI map layer shows the 4-corner point, but the county layer doesn't match and has an offset. I'm not sure I would know which to use without meets and bounds. In any case we agree that there would be no connection at that point.

[muon2]

Dunn, NC has its own micropolitan statistical area, because it is the largest town in Harnett, County, and about three times larger than the county seat of Lillington. Lillington is the fourth largest town in the county. This is a reason not to concentrate on the county seat.

I actually use the seat of government as the node of a political unit. Since the counties represent political jurisdictions, the seat of government gives a more precise point than a municipality within the county. This definition then naturally scales to the seats of government for subunits of a county. Only when a chop amalgamates multiple subunits do I look to the most populous of those units to define the node, and then I look to the municipal building for that largest subunit.

Its important to note that the seat of government is not always in the county seat. A number of counties have built county offices out side the municipal county seat when larger space is needed, and that can matter in determining the shortest path. In St John the Baptist Parish LA, Edgard is the county seat and the main courthouse is there, but the county government meets in LaPlace so that is the node. Those towns are on opposite sides of the Mississippi and there is no connection within the county, so the distinction matters in determining connections.

[muon2]

I thought my definitions were clear that I do not use the courthouse, I do measure the node from the building where the government meets.

-----

As I was working on CO, I found an interesting puzzler. Consider Pitkin county, home to Aspen. I'm familiar with it and the roads I'll describe having vacationed and had conferences there a number of summers.

To the east is Lake county connected by CO-82 over Independence Pass. At over 12,000 ft it's the highest paved pass of the Continental Divide in the US, but it's closed in the winter.

To the north is Eagle county. CO-82 crosses the SW corner of Eagle from Pitkin before going into Garfield. From that little segment one can take local roads including unpaved stretches through high (and stunning IMO) alpine meadows of the Sawatch range to connect with I-70 and the county seat of Eagle. So they are locally connected, but only by unpaved roads.

To the northeast is Garfield county and Glenwood Springs is the county seat one would most likely reach in the winter from Aspen along CO-82. As I mentioned that cuts about 3 miles through Eagle county between Pitkin and Garfield. You can get from Pitkin to Garfield on back roads that skirt the north side of the Elk mountains, but that includes unpaved roads, too.

To the west is Mesa county, but the short border is made up of the high peaks of the Elk mountains and there are no connecting roads.

To the south is Gunnison county. CO-133 goes from Pitkin to Gunnison, but there are no connections (paved or unpaved) from the rest of Pitkin to CO-133 without going into Garfield (or Eagle and Garfield if paved). Even so, CO-133 doesn't connect to the Gunnison county seat except over unpaved local roads. To the SE there are local crossings from Pitkin to Gunnison, but they are unpaved and closed in winter.

Pitkin may be such an exceptional case that it isn't worth adjusting rules for, but I thought it interesting enough to include in the thread.

[muon2]

Road connections should be all weather roads in my opinion (they're snow plowed). So using the road over Tioga Pass in Yosemite is a no go in my opinion.

You were clear about the seat of government. I had a reading comprehension problem. I still think the county seat however should be used for the reasons I stated.

That's the problem with Pitkin. There are no all-year paved roads connecting Aspen to another county seat without passing through an intervening county. It works for jimrtex since he requires only contiguity and a direct connection that can pass through an unconnected county.

I would be inclined to treat Pitkin as a special case, due to the lack of any other alternative. There is an unpaved connection to Garfield that includes low priority plowing, but at least it is something more than just closed in winter.

[muon2]

There is often more than one valid path between county seats. If a county is chopped it matters which path is the connecting link. That requires selecting a particular point from which to measure the paths. That led me to select the state highway nearest the seat of government building as that point.

The alternative to selecting a preferred path is to automatically create links from a neighboring county to all pieces of a chop if there is a contiguous border with any regional (or local) connection. When we first looked at this a couple of years ago with OH as our subject, the extra links tended to generate too much erosity around rural chopped counties.

[muon2]

So, Torie it sounds like you would favor a combination of the blue, green and gold links on my map below. That would exclude the local connections (yellow) where there is no nearby highway to establish a link. You thought that jimrtex's definition of a direct route was unworkable. Do you have a definition of how much corner one can cut and still count?

In this map I have revisited the yellow links on jimrtex's map.

Blue are regional connections based on continuous state highways between two county seats that don't enter a third county. The dark blue links were those not identified previously, but meet my criteria.

Green is a all year ferry connection that meets the criteria for a regional connection.

Yellow are local connections that rely on local roads to establish a path between counties.

Orange are near connections based on state highways where the highway path cuts a short distance through a third county, such as at a corner.

Gold are connections equivalent to both yellow and orange.

Pink are contiguous counties without a connection.

Red squares are places with point contiguity.

[muon2]

Precincts are tricky since they can change at the whim of the county. You could use the definition for an urban area from the Census:

You must be logged in to read this quote.

Maps of all urban areas are maintained by the Census Bureau so it is straightforward to check if a direct road passes through one.

[muon2]

You be the judge. Here is a link to all the maps of all the urban areas in the US. The natural thing to do is check the maps of NC munis in the vicinity of all the yellow and gold lines on my map. They are locally connected counties, so one just has to see if a route through other counties avoids any of those areas.

[muon2]

In the example immediately above the local roads provide an internal path should only those counties be in a district together. I think that Torie agrees that when a district is complete one should be able to drive to any point in the district without leaving the district. The only exception is when an area is is isolated but part of a whole geographic unit. This would apply to cases like the parts of St John the Baptist parish on either side of the Mississippi. They are not locally connected, but can be together in a district if the whole county is in the district.

[muon2]

To clarify I had lengthy conversations during the 2011 remap with experts hired in the past by the controlling party. One repeated comment they made was that contiguity without connection (rivers, railroads, etc) allowed them to design some of their best gerrymanders. So, they suggested that one of the best rules one could enact to curb the power of gerrymanders was to deny them that ability.

[muon2]

To clarify I had lengthy conversations during the 2011 remap with experts hired in the past by the controlling party. One repeated comment they made was that contiguity without connection (rivers, railroads, etc) allowed them to design some of their best gerrymanders. So, they suggested that one of the best rules one could enact to curb the power of gerrymanders was to deny them that ability.

Were they using building blocks such as counties at large scale, with townships if necessary to split counties?

My connection between Cleveland and Akron is comprised of whole census blocks and contains not mere paved streets, but interstate highways.  I even took care to include the parts of a cloverleaf so you wouldn't have to leave the district driving in either direction.

They used whatever they could get away with, since gerrymandering was their goal.

[muon2]

I'm looking at the local connections on my NC map. I'll start with the western end. If there is a modification to the connection definition, we should see how it plays in the mountains.

Cherokee (Murphy) to Macon (Franklin)
The direct road is US 64 which is 49 mi and spends 27 mi in Clay. There are no urban areas in Clay and US 64 has a bypass around the Clay seat (Hayesville). Cherokee and Macon are contiguous and US 19/US 74 crosses the border, but only travels along the edge of Macon on its way north. To get to Franklin requires considerable travel on a winding local road (Wayah Rd or Junaluska then Wayah), so it's locally connected. This failed both the jimrtex and muon2 criteria.

Cherokee (Murphy) to Graham (Robbinsville)
The direct road is US 129. Some renderings of US 129 have it clip the corner of Macon (Bing) while others show is straddling the county line (Mapquest). The satellite view shows it on the county line. I allow roads on the line to count in either county. This passes both the jimrtex and muon2 criteria.

Graham (Robbinsville) to Macon (Franklin)
The direct road is NC 143 and NC 128 which is 45 mi and spends 14 mi in Swain. The route does not pass through any urban areas and doesn't approach the Swain county seat (Bryson City), though it does follow part of the preferred route one would take to get there from Graham. Graham and Macon are contiguous and a US highway crosses the border. One can use the the aforementioned local Wayah Rd (it's one of the NC secondary roads), so it's locally connected. This failed both the jimrtex and muon2 criteria.

[muon2]

Graham (Robbinsville) to Macon (Franklin)
The direct road is NC 143 and NC 128 which is 45 mi and spends 14 mi in Swain. The route does not pass through any urban areas and doesn't approach the Swain county seat (Bryson City), though it does follow part of the preferred route one would take to get there from Graham. Graham and Macon are contiguous and a US highway crosses the border. One can use the the aforementioned local Wayah Rd (it's one of the NC secondary roads), so it's locally connected. This failed both the jimrtex and muon2 criteria.

You are most of the way from Robbinsville to Bryson City before you turn and go south. This is the route to Swain County, even NC 128 continues south.

That's what I assumed based on your definition of direct route, though you didn't say how much of the initial portion must be in common. Presumably using a main road to get out of the originating town is not enough to be considered the initial portion. I imagined that you would apply something like 1/3 of the route, consistent with the (2)(b)(i).

Direct Route: A route is direct if:
    (1) it it is not circuitous; and
    (2)(a) it is entirely within the two terminating counties; or
        (b)(i) less than 1/3 route is within the intervening county (-ies); and
             (ii) the initial portion of the route would not form part of the quickest route to more than   
                  1/3 of the residents of the intervening county. This calculation should be done in both
                  directions.

We agree that it shouldn't count as a connection for different reasons. Does Torie want a different outcome?

[muon2]

I like Bing maps for its ability to drag not only the route, but also the start and end points. Bing also shows the county lines crossed along on the route. I used that to first find the route and distance from Robbinsville to Franklin, then I dragged the ends to the Swain county lines along the route to get that segment length.

I haven't dismissed the idea of local connection plus nicked direct connection as an alternative. I just want some specific language I can test both for whole county connections and when counties are chopped.

Here's where I think the Torie plan is at present, so correct me if I'm wrong.

Two counties are regionally connected if they are locally connected and have either

  A) a continuous path of numbered state highways between nodes completely within the two counties; or

  B) a continuous path of numbered state highways between nodes that includes a segment outside the two counties provided that the segment
    1) is on the most direct route between the nodes; and
    2) comprises no more than one third of the most direct route between the nodes; and
    3) passes through no Census-defined urban areas; and
    4) carries the same state or federal number on its length.

[muon2]

DRA uses Bing maps as its source. When I look at Benton to White I see the state highway symbol for IN-18 on Division Rd (Benton). When I have doubts about whether the number is part of the state system, I find the Wikipedia is usually up to date on state highways (cf Indiana State Road 18) and includes a nice table showing the counties and major intersections. For close calls at corners I flip between Bing road and satellite images. In any case 18 is a connection.

One issue with lines in DRA is they represent things as they were in 2010. That's good for our exercise redoing the past, but may not be accurate if we are looking to 2020.

[muon2]

I would never have drawn my first map, if using the old Indianapolis city lines was a no go. Locality chops in a macro-chopped county are almost always a scoring disaster. Your metrics give one a great incentive to avoid that. And presumably that is justified on the theory that it will tend to avoid more substantial partisan variations between maps which are competitive on the pareto optimal frontier.



Do you agree with my road cut count for my artistic endeavor (65 cuts)? Did I do it right? Did I do the count right for the chopped county of Henry? Is there an extra cut when an applicable state highway goes in and then out again of the chop before going into another county (which is what I did)?

The Hendricks to Marion/Pike connection is one that I wouldn't count for erosity. There is a large reservoir and park that generally block east-west travel. I-74 cuts a corner of Pike but has no interchange in the township. I-65 cuts a corner of Hendricks but has no interchange in the county. 56th street is the main through road across the water into Pike but isn't a state highway.

The two are locally connected, but based on the Kent analysis the severed link shouldn't count for erosity (yes one of will have to search for the thread at some point to remind the readers of the work we did. )

[muon2]

"56th street is the main through road across the water into Pike but isn't a state highway."

You only count state highway cuts for macro-chopped county quasi counties?

Subunits within a county are connected if they are locally connected. Subunits are connected across county lines if they are regionally connected. This was the Kent determination confirmed by application to the Detroit UCC. If we counted all local connections, the erosity along a straight county line got too high due to the number of broken links.

[muon2]

Here are some additional links that are missing above.
Elkhart to Noble at US 33/US 6
Clinton to Madison at IN-18
Madison to Delaware at IN-32
Putnam to Morgan at IN-42
Parke to Vigo at US-41
Dubois to Orange at IN-56
Orange to Crawford at IN-37
Crawford to Washington at IN-56
Rush to Decatur at IN-3
Interestingly these are all bright red lines on my trusty paper Rand McNally Atlas.
With the previous adjustments, that makes an erosity score of 73 with a chop score of 3. What is your population range by the way?

I think this has all the county links with my definition. It can serve as a starting point to test other definitions.

[muon2]

Frankly, color me skeptical that ferry routes are the best ways to connect unconnected areas--it seems to me that they should only be used for islands not connected to the mainland by highway, and the like.

The ferry rules came about in our WA discussion along with the mountain crossings. We felt that Puget Sound crossings such as King to Kitsap, Snohomish to Kitsap, or Island to Jefferson should be allowed. Island is connected northeast to Skagit by bridge, but there is a lot of traffic on the ferry to the southwest as well (confirmed by my travel through the county this summer).

[muon2]

Yeah speaking as someone who lives in Whatcom County, crossing the Cascades in the North is a no-no.  The big issue with mapping Washington north of Seattle is that Whatcom+Skagit+San Juan+Island (which together form a reasonable community with lots of commuting between those regions and such) are nowhere near close enough to a full congressional district on their own and require chopping up Snohomish County, which is bad because that region is essentially just an extension of the Seattle suburbs.  Ferry crossings are weak too imo but should be acceptable if it means crossing the Cascades as little as possible and I think are understandable if not screwing up the Puget Sound area.

The biggest problems with the current map that would need solving in any sort of non-partisan redistricting are:

1. Getting rid of my district, the hideous 1st, which has no road connectivity and is essentially just the product of a totally unnecessary Larsenmander.  

2. That idiotic three-way split of Tacoma, which itself is a product of

3. The so-called majority-minority district, which acts as nothing but PR for the WA Dems, as none of the minority groups in the district are realistically large enough to elect a representative of their choice.  It's effectively just a Democrat pack which totally screws up the Puget Sound region on the map.

Ideally I'd also like to see a map which limits the Cascade crossings to one district, preferably along the Columbia, though from my own mapmaking attempts I've realized that this results in chopping up Yakima.  Putting Kittitas and Chelan in a King County district is imo both ridiculous and historically unprecedented and worth the Yakima chop but I understand that the metrics you guys work with will disagree.

I agree that WA-1 is a mess by our metrics, though it is admittedly compact geographically. It's a good example of the shortcomings of traditional compactness.

There's an additional problem with the Yakima chop. It has to slice up the metro area to get enough population. That leaves the King crossing.

[muon2]

Wait, is Yakima a UCC of its own? It's not colored on the stickied Jimrtex map.

In any case, that raises the eternal question of whether chopping a single county UCC should extra-penalized relative to a normal county chop...

The consensus has been to not double penalize chops in single county clusters.

[muon2]

County lines may be arbitrary., but most people know the county and municipality they live in. That makes county a CoI even if not the strongest one. Counties have the huge advantage of virtually never changing boundaries. That makes whole counties relatively resistant to gerrymandering. Note that states with criteria for independent mapping generally seek to minimize split counties.

[muon2]

Metrics aside, #3 is definitely the best map.  The chops in Grays Harbor and Lewis look a little fugly, and District 8 makes me uncomfortable, but it's worth it because that's probably the prettiest Puget Sound I've seen in any Atlas redistricting attempt.  

All in all, nice work.

Thank you. The beauty of the system, as I said before, is that it is all automatic. Partisan bias is totally exorcised. And the temptation to like a map that favors one's party is overwhelming. Almost none of us can resist it. I am getting better at resisting it, because I am about in equipoise between the two parties at present. I disdain them both. My agenda is more swing CD's. I want more centrists in Congress. But the system favors nobody's agenda, other than making the job easier for election boards, because there will be far fewer chops of governmental jurisdictions, and ballot preparation will be a heck of a lot easier.

Come to think of it, map 3 involves a bridge chop (Pierce and Kittitas are not road connected by any pavement or vehicular ferry whatsoever so in my scheme not deemed connected for any purpose), so it loses to map 2 as the tie breaker before even reaching erosity scores, unless map 3 avoided a locality chop that map 2 did not, which it does not in this case. In my scheme, as a tie breaker in this sense, while bridge chops are allowed, they are disfavored.

Oh, and one final thought. Given all the barriers in Washington, I suspect this state has one of the highest potential partisan swing effects of one map versus another which are competitive per the metrics. Small changes in population, can move the map around involving moving tons of folks from one district to another. That is because the barriers basically force such changes. One cannot jiggle the lines a bit from county to county, because they are both outside the east, quite large, and have various barriers. So in that sense, a SKEW swing of from 1 to 3 is reassuring, albeit that is a huge swing for such a relatively small state. If the state where three times a big a swing of 6 seats rather than 2, would be rather frightening. But such a large state with such barriers does not exist.

There's no bridge chop in map 3 that I see. A bridge chop links two whole counties. Pierce is chopped so the bridge rule doesn't apply. Pierce already gets a chop penalty so it doesn't need to be impacted by the bridge rule from King. If both King and Pierce shared chops with anther district you'd have a traveling chop, but that doesn't apply either.

[muon2]

Metrics aside, #3 is definitely the best map.  The chops in Grays Harbor and Lewis look a little fugly, and District 8 makes me uncomfortable, but it's worth it because that's probably the prettiest Puget Sound I've seen in any Atlas redistricting attempt.  

All in all, nice work.

Thank you. The beauty of the system, as I said before, is that it is all automatic. Partisan bias is totally exorcised. And the temptation to like a map that favors one's party is overwhelming. Almost none of us can resist it. I am getting better at resisting it, because I am about in equipoise between the two parties at present. I disdain them both. My agenda is more swing CD's. I want more centrists in Congress. But the system favors nobody's agenda, other than making the job easier for election boards, because there will be far fewer chops of governmental jurisdictions, and ballot preparation will be a heck of a lot easier.

Come to think of it, map 3 involves a bridge chop (Pierce and Kittitas are not road connected by any pavement or vehicular ferry whatsoever so in my scheme not deemed connected for any purpose), so it loses to map 2 as the tie breaker before even reaching erosity scores, unless map 3 avoided a locality chop that map 2 did not, which it does not in this case. In my scheme, as a tie breaker in this sense, while bridge chops are allowed, they are disfavored.

Oh, and one final thought. Given all the barriers in Washington, I suspect this state has one of the highest potential partisan swing effects of one map versus another which are competitive per the metrics. Small changes in population, can move the map around involving moving tons of folks from one district to another. That is because the barriers basically force such changes. One cannot jiggle the lines a bit from county to county, because they are both outside the east, quite large, and have various barriers. So in that sense, a SKEW swing of from 1 to 3 is reassuring, albeit that is a huge swing for such a relatively small state. If the state where three times a big a swing of 6 seats rather than 2, would be rather frightening. But such a large state with such barriers does not exist.

There's no bridge chop in map 3 that I see. A bridge chop links two whole counties. Pierce is chopped so the bridge rule doesn't apply. Pierce already gets a chop penalty so it doesn't need to be impacted by the bridge rule from King. If both King and Pierce shared chops with anther district you'd have a traveling chop, but that doesn't apply either.

WA-08 goes from the whole county of Kittitas to take part of King and then going to Pierce. That's a classic bridge chop to me unless Pierce and Kittitas were themselves adjacent, and without any qualifying pavement connection, they are not. That is how I see it anyhow.

A bridge chop only applies to a chopped county used to connect two whole counties. The rule is to prevent a district from bypassing an urban center in the middle of the district by chopping out the center, but otherwise keeping counties whole. It often creates the worst of the dumbbell shape that you have taken exception to in the past. Connecting a sequence of chopped counties, as long as it isn't a travelling chop is ok.

Pierce isn't whole so there is no bridge chop. Pierce and King share no other district so it isn't a travelling chop.

[muon2]

I found my notes from late 2012 when we were looking at the question of regional connections for WA. This was the connection map I found, with numbers equal to the percent of a CD. The red lines partition the state into whole county groups equal to a whole number of CDs. The purple and orange lines show two different partitions I was looking at.



From that I reconstructed the individual CDs. There are only 6 county chops plus the pack penalty in Seattle for a CHOP of 7. CD 4 is erose, but not quite as bad as it looks because of the mountains. The political breakdown is 4D, 2d, 2e, 2R for a SKEW of 2.





Within the Seattle area Vashon only connects by ferry to Seattle within King. The chop into Snohomish is designed to provide a connection to the Stevens Pass portion of King.