Talk:Tag:place=neighbourhood

Boundaries

putting this forth for information and discussion - and to keep me consistent

There was some discussion during the proposal period of using admin_level=* in conjunction with a boundary way used to map a subdivision with defined borders. As pointed out on the main page, subdivisions have a variety of statuses depending upon place. In my particular case, I am mapping in Delaware, United States and subdivisions are included in the county government GIS data, under the administrative care of the Department of Land Use (see http://www.nccde.org/parcel/search/ for search interface). In my case, I'm thinking that admin_level would not be apropos, but rather use of border_type=subdivision. I'm representing the subdivision itself as a relation:boundary:

For the relation:

border_type {subdivision}

name {Mayfield}

type {boundary}

wikipedia {en:Mayfield, Delaware}

For the boundary way itself:

boundary {administrative}

source {New Castle County eParcel Database}

source_ref {http://gismaps.nccde.org/NCCPortal/searchPage.jsf}

source:date _use the date on which the database was consulted_

(role = outer in relation)

(where a boundary is contiguous with a roadway, place either on road edge or shoulder; don't overlap with road way)

For the label node:

(various previously imported data)

name {Mayfield}

place {neighbourhood}

(role = label in relation)

--Ceyockey 15:47, 31 December 2012 (UTC)

Mapping as an Area or Relation vs Node

I've updated the page with further discussion about the cases when a node or area is appropriate and useful.

Also see Key:place#Areas and Talk:Key:place for further discussion of tagging places as nodes vs areas, in particular: Talk:Key:place#Nodes_and_areas_for_settlements and Talk:Key:place#Verifiability_of_mapping_populated_places_as_areas --Jeisenbe (talk) 02:44, 27 April 2019 (UTC)

Review of usage on areas

This tag is not used consistently on areas. Usually when it is done, it is duplicating a node, a landuse=residential feature, or an administrative boundary.

Singapore:

way 474215388 - duplicates node 4688137655 - all others (30) are mapped only as nodes. Appears to arbitrarilly exclude the adajacent residential area (separated only by a highway=service): way 172471342

Costa Rica:

5228 nodes(!), 13 ways, 1 relation: http://overpass-turbo.eu/s/RQZ

Bremen:

No ways or relations, 90 nodes: http://overpass-turbo.eu/s/RQU

Kaliningrad:

No ways or relations, 24 nodes: http://overpass-turbo.eu/s/RQV

Northern Ireland:

4 nodes (all named), 1 way: https://www.openstreetmap.org/way/409244094 - 200m x 100m, unnammed!

Karlovarsky, Poland

12 nodes, no ways or relations: http://overpass-turbo.eu/s/RQX

Latvia

19 nodes, 2 way, 2 relations: http://overpass-turbo.eu/s/RQY Ways: way 307852629 - also landuse=residential; way 156980758 - also barrier=fence, gated service road (private) with only 8 houses - should be landuse=residential (gated community). Relations: both are type=multipolygon, one is also landuse=allotments (relation 1762429), the other is also mapped as a node (relation 4744483, node 3379311587

Corsica

16 nodes, 4 ways: the ways are all in 2 adjacent villages, added on the same day in 2016.Way 434979743 includes only 2 houses and is a rough oval. Way 435115346 includes 6 buildings, also round, etc.

Zeeland, Netherlands

84 nodes 3 ways: way 6318482: tagged with landuse=industrial(!). And these two, which are almost adjacent but dont' share nodes, rather having a random gap in between: way 416688755 and way 416549074. 15 boundary=place relations. All are duplicating nodes, all are in one town: node 3123884752 - Terneuzen.

Australia Capital Territory

5 nodes: http://overpass-turbo.eu/s/RR0

Uruguay

371 nodes 46 ways - 23 are small areas with landuse=residential (4 duplicate a node), 50 relations - 30 with landuse=residential, 20 with boundary=administrative. I didn't check how many are also nodes.

Bali, Indonesia

38 nodes only.

Hawaii, USA

72 nodes, 47 ways, 2 relations - so that's where they are at. (This is by overpass; in my older sample from October there are more ways and fewer nodes:) Ways: 45 are landuse=residential (and most of these look appropriate for that tag: small areas, just a few block, purely residential, 1 entrance road), 1 is landuse=farmland (but should be residential it appears), and 3 are nothing else but a name: two have the same name but are separeted? (way 589996026 and way 589996025). Relations: One is an admin_level=9 boundary and mainly includes uninhabited hills, but also 10(!) place=neighborhood nodes: relation 10045916. The other is a boundary=census, tiger:reviewed=no relation 119260.

Delaware, USA

133 nodes, 14 ways, 5 relations (same in overpass and in my pbf from october) Ways: 2 are landuse=residential, 9 are boundary=protected_area?! - these are official "historic districts" but the shapes are very jagged, suggesting that the named place is not the same as the historic district. All are in central Wilmington city, except for one which is part of Newcastle village/town. The other 3 are all next to each other, in southwest Newark town, and look like landuse=residential areas (named subdivisions). Relations: All 5 are imported from "New Castle County eParcel View Map". Each one is mapped with a node, plus the highway=residential ways nearby! They are type=site - quite odd.

Prince Edward Island, Canada

12 nodes, no ways or relations: http://overpass-turbo.eu/s/RR1

Tasmania, Australia

159 nodes, no ways or relations: http://overpass-turbo.eu/s/RR3

Summary

This review suggests that few neighborhoods are mapped as areas, and in most cases this mapping is low-quality and the area chosen is not verifiable, except if it duplicates an administrative boundary or landuse=residential area. I would recommend only using place=neighbourhood on nodes, as with other settled places.

If there is a verifiable administrative boundary, or a small fenced/gated residential area, these can be mapped with boundary=administrative relations, or landuse=residential areas. --Jeisenbe (talk) 05:29, 24 March 2020 (UTC)

is_in

As

• is_in=* is actively removed by JOSM, and as far as I know there is no opposition to that
• nowadays typical administrative boundary is mapped as an area (or maybe it is true only in Europe and USA)?

I think that is_in=* should not be listed as recommended combination Mateusz Konieczny (talk) 09:05, 2 December 2020 (UTC)

The tag is_in=* is "recommended" on a lot of pages from when it was used. It is not actually deprecated, but is an unnecessary tag as geospatial systems are smarter now.

I've gone ahead and removed it. --GoodClover (talk) 15:33, 2 December 2020 (UTC)

This is really only true for Europe and the richer countries of East Asia. In many places the admin boundaries do not correspond well with named places. In the USA there are large towns which are not municipalities, so have no enclosing boundary below admin_level=6 (the "county"). In Indonesia most of the high-number admin_level features, such as towns and villages, are not mapped and there is often no good official source for the boundaries, so it is quite difficult to add them. --Jeisenbe (talk) 04:08, 3 December 2020 (UTC)

The tag is_in=* has been used with ~5% of all place=neighborhood features, a rather high percentage. Since most are mapped as nodes, it is a way to associate the neighborhood with the place=city or place=town to which it belongs - which are also mostly mapped as nodes and may often not correspond to an admini area. --Jeisenbe (talk) 04:11, 3 December 2020 (UTC)

Rendering

Is this rendered on any map styles currently? If so, would be good to document status.

It is rendered by OSM Carto (zoom levels 15 to 19). Yes, it would be good to document this; for example, city_block is not rendered, but you wouldn't know this. But there doesn't seem to be a way to simply add "Rendered by OSM Carto: Yes/No" to the info box. Phodgkin (talk) 11:36, 11 February 2023 (UTC)