[Talk-us-newyork] New York minor civil divisions - status and progress

Wed Aug 24 17:11:30 UTC 2022

On Wed, Aug 24, 2022 at 12:05 PM D. Joe <osm+joe at etrumeus.com> wrote:

> On Tue, Aug 23, 2022 at 09:56:27PM -0400, Kevin Kenny wrote:
> >    At long last, I'm reasonably satisfied with the status of the
> relations
> >    representing the minor civil divisions (admin_levels 7 and 8) in New
> >    York State.  All 1600 or so have been compared against NYSGIS and
> >    TIGER/Line 2021, with a great many boundaries in OSM redrawn because
> >    the data were an unholy mess when I started.
>
> Awesome. Thanks for that.
>
> I don't suppose school district boundaries are part of this, by any chance?
>

All special assessment districts (school, fire, water, sewer, library,
etc.) were out of scope.

> On another scale entirely, do we have relations for the regions, as
> defined initially by Empire State Development but which came to prominent
> use during the pandemic:
>
> https://esd.ny.gov/regions
>
> (I understand if this wasn't in-scope for your recent work but I figure
> I'd ask just to be clear.)
>

Again out of scope. The boundaries of these regions have not been
historically stable. Until they started being used as groupings for
CoViD-19 statistics, they were generally just carved up based on page count
in tourism guides. If I wanted to map them, I think I'd create
superrelations comprising the counties that make them up, since they all
follow county lines.

> [...]
> >    I have questions remaining about how best to do some of the
> >    finishing touches, particularly relating to the `label` and
> >    `admin_centre` nodes.  Rather than burden the list with a lengthy
> >    explanation, I've dropped an OSM diary entry at
> >    [1]https://www.openstreetmap.org/user/ke9tv/diary/399813 (Replying on
> >    either the mailing list or in diary comments is fine.)
>
> I'm deeply uneasy at the thought of a proposed scheme for tagging places
> in New York based on arbitrary population threshholds, while in the process
> overloading and obscuring how those terms are used every day in the state,
> eg, city, town, village, hamlet.
>
> Maybe I misunderstand what's being proposed here, and why? Maybe the
> meaning of the terms in the sense already used locally is carried by a
> different tag? Or maybe the population information could be carried by a
> different tag?

I think it's all in there.  Compare Village of Pittsford
https://www.openstreetmap.org/relation/176068 and Town of Pittsford
https://www.openstreetmap.org/relation/2763735. Both are clearly identified
(look at `border_type` on the relations, too)  and at the correct
`admin_levels`.  A single `place` kind of has to do for both, since the
village truly is the administrative and cultural center of the town. They
have their distinct boundaries, and both their populations are tabulated.

`border_type=*` tagged on the boundary relation identifies County, City,
Town, Village and Hamlet.  Because different regions of the same often
abut, `City of`, `Town of` and `Village of` are retained on borders (but
not on `place=*` nodes). Fear not, City of Tonawanda and Town of Tonawanda
retain separate identities!

Population on boundaries is always the population of the enclosed region,
so every County, City, Town, Village has its population tabulated, as do
the subset of Hamlets that are also CDP's. The population tag on the place
node is a rendering convenience, and is the population of the smallest
region that the node labels, to avoid giving, for instance, the village of
Pittsford a label based on the whole township.

I understand the desire to convey some notion as to the population an
> "area" might have. When I'm learning about a place in the US I'll look at
> whatever the nominally urbanized population center there might be, but I
> also tend to calibrate my sense and expectations as to how "big" a "place"
> is based on the population of the relevant federal statistical area, eg,
> Rochester at only about 200K people is the core of a much larger, eponymous
> Metropolitan Statistical Area (about 1M people).
>

This breakdown isn't intended as a hard-and-fast rule.  It's a starting
point. It turns out that when I looked at the boundary cases, there were
only a few that I'd tweak because they sorted out pretty well in any case.
I'd probably promote Binhgamton and Niagara Falls to cities, for instance,
and definitely want to retain Saranac Lake as a town even though it's tiny
(the only hospital and airport for many miles around is a key point there).

Simply choosing a Metropolitan (or Micropolitan) Statistical Area also
wouldn't be a perfect fit, either.  Albany/Schenectady/Troy/Saratoga
Springs are a single MSA, but we wouldn't want to glom them together for
mapping purposes. The MSA's also change from one Census to the next, while
political boundaries are considerably more stable.

> Considering the towns in Monroe County surrounding but outside of
> Rochester proper, it is rare, but there are a couple of instances where
> there is a village within a town of the same name (there is a village of
> Pittsford in the town of Pittsford; a village of Webster in the town of
> Webster). In each case, the village has a distinct government from the town.

It's not rare at all in the state.  Both communities are mapped, with
populations attached to their borders.

In the case of Villages, the Village is subordinate to the Town (the Town
continues to provide some services and the residents pay taxes to both), so
Villages are at `admin_level=8` while Towns are at `admin_level=7`.  The
case of a City that has been carved out  of a Town of the same name is a
little harder, and is the chief reason for keeping 'City of' and 'Town of'
in names - a border between Tonawanda and Tonawanda, or Plattsburgh and
Plattsburgh, looks exceedingly weird.

None among Irondequoit, Penfield, Brighton, Henrietta, Chili, Gates nor
> Greece have an eponymous village or hamlet or indeed any smaller urbanized,
> named, municipality with its own government. Wikipedia tells me East
> Rochester is both village and town (coterminous, and so a distinct
> situation from those in Webster and Pittsford).
>

Consolidated towns and villages have a single set of border ways.  They
have boundary relations at both admin levels. The same is done for the five
boroughs of New York City, which are mapped as both county
(`admin_level=6`) and `suburb` (in the OSM sense of a named division of a
larger city) at `admin_level=7`.These last need the separate handling
because three of the five have different names as county and as borough
(Borough of Brooklyn = Kings County; Borough of Manhattan = New York
County; Borough of Staten Island = Richmond County).  Then again, the
handling here is always going to be a bit odd, because New York City is the
only `admin_level=5` in the US - it's the only city that has consolidated
multiple counties. Which, contrary to what people have asserted here, still
exist as administrative entities.  While they have ceded all legislative
functions to the city, they retain a (largely advisory and ceremonial)
executive branch, and have kept an independent judiciary; every county in
New York City has a county court.

> Further out, the reverse is true: It's unusual for there not to be a
> village within the outlying towns of the county, but in each case the
> village is named distinctly from the town which encompasses it: Fairport in
> Perinton, Honeoye Falls in Mendon, Scottsville in Wheatland, Churchville in
> Riga, Spencerport in Ogden, Hilton in Parma, and Brockport in Sweden. Only
> the outlying towns of Rush, Clarkson and Hamlin lack a smaller, more
> urbanized municipality within themselves.

That's where the arbitrary `place=region` (soon to be `place=municipality`)
comes in. Town of Perinton gets a label, Village of Fairport gets a label.
I'd have to look up what Perinton got for an admin centre.

> I'm looking at
>
>
> https://commons.wikimedia.org/wiki/File:Monroe_County_(New_York)_-_Towns,_Villages,_and_City.svg
>
> and it reminds me--did you tangle with the convolutions of the
> Rochester/Irondequoit boundary? Rochester apparently has two narrow spurs
> of territory, one leading to Durand Eastman Park and the other leading out
> to Irondequoit Bay (to accomodate a storm water system). Both spurs cut
> across Irondequoit, leaving it in in three formally disconnected pieces
> (but not entirely an enclave/exclave situation, since each portion of
> Irondequoit is not entirely surrounded by Rochester, bordering as it does
> either Lake Ontario or Irondequoit Bay).
>

I did, and it's indeed a tangle!
https://www.openstreetmap.org/relation/176069

Rochester encompasses most but not all of Durand Eastman Park (there's a
region to the southwest that's in Irondequoit). It owns the rights-of-way
of Culver, Hoffman and Wisner Roads connecting up to it  It also owns the
flood plain of Dunsmore Creek. The border is also a tangle to the
southwest. Rochester owns Genesee Valley  Park and the airport, but Chili
owns most of the surrounding area (including Genesee Valley Greenway and
the canal right-of-way), making the airport a true exclave.  Rochester also
has another narrow strip sticking into the Town of Greece through the
middle of Eastman Business Park, apparently also for stormwater management
since there's a creek in it. Irondequoit is mapped as multiple polygons
since in all the data I have, the townships stop at the shoreline. Cities
can be a different case: the City of Glen Cove is mostly water and extends
to the middle of Long Island Sound. New York City's boroughs have
well-defined borders that follow Federal bulkhead lines (rather than the
shoreline) as they stood in 1898.

-- 
73 de ke9tv/2, Kevin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us-newyork/attachments/20220824/a6344efa/attachment-0001.htm>