From OpenStreetMap Wiki
Jump to navigation Jump to search

Neighborhood vs Subdivision

For the purposes of this discussion I propose we use the term neighborhood and subdivision. I'm not sure if it's warranted to create two tags for this purpose, but some may believe it is, and it's better to have just the term subdivision instead of mixing in the equivalent development, and wordy residential community. -- Joshdoe 13:37, 14 April 2011 (BST)

I would vote against using more than one value for whatever key we agree upon (see below). Especially if you look at the global level, the difference between a neighborhood and a subdivision is minimal. Both basically boil down to an area defined by name either through its history or by a development-project. -- OliverH 14 Apr 2011 16:52:00 +0200
I'd agree, there's no real difference except one has a well defined border. In terms of data consumers it likely doesn't matter, or perhaps someone could add a tag down the road to indicate it is a planned neighborhood (subdivision). -- Joshdoe 16:41, 14 April 2011 (BST)

Node vs Way

As mentioned on the page, neighborhoods don't always have well defined boundaries, while subdivisions nearly always do. In other words all subdivisions can be represented by areas (though can be mapped as nodes), but not all neighborhoods can be mapped as areas (though you should try, and use a node otherwise). -- Joshdoe 13:37, 14 April 2011 (BST)

Choice of the key

If using place=* is the only way that data consumers such as nominatim can account for the neighborhoods I think we can quickly settle on using that as the key. The tag can be used both on areas (clearly defined) and nodes (not so clearly defined). -- OliverH 14 Apr 2011 16:52:00 +0200

I'd agree, place=* makes sense to me. All we have to decide upon is a value, whether subdivision, neighborhood, neighbourhood or whatever. -- Joshdoe 16:41, 14 April 2011 (BST)
I think that place=* is ok but only if it is intended to be used for point geometries mainly. There's no tradition of place=* with polygons (a place=city on a polygon would become boundary=administrative with an appropriate admin_level). --Frederik Ramm 10:15, 1 September 2011 (BST)

Neighborhoods within neighborhoods

In my local area we have named subdivisions which I would tag as a neighborhood, but they also belong to larger named areas which I would also consider neighborhoods. It would be nice if there were some way to indicate a "sub-neighborhood" is part of a larger "neighborhood". We could use boundary=administrative and admin_level=* (10/11/12/etc, dependent upon country/state/etc) for this purpose, though I know some would consider it a misuse of the tag since it's not a governmental boundary. It appears Germany and Netherlands uses admin_level=11 for neighborhoods (see here), however perhaps neighborhoods are actually administrative areas there. -- Joshdoe 16:41, 14 April 2011 (BST)

admin_level 11 ends at borougs of about 10.000 as far as I (as a Dutchman) can tell. After that there is room for smaller areas that I'd consider neighborhoods (anywhere from 5 to a couple thousands of inhabitants) I'm planning to adress my issues on the NL mailinglist (which is the reason I found this discussion so soon after you created the page). --OliverH 14 Apr 2011 22:17 +0200

Neighbourhoods vs suburbs

Why do we need a different key from suburb? place=suburb is used for various administrative levels, and often nested. Can we articulate a real defined or quantitative difference between a “suburb” and “neighbourhood?”

Here in Winnipeg, the closest thing to a suburb is the set of former towns and rural municipalities which joined to form the city. Roughly corresponding to them we have community areas used for regional health planning, divided into neighbourhood clusters, further subdivided into neighbourhoods. Politically, we have community committees, divided into wards (each with a seat on council), further subdivided into the same set of neighbourhoods. The community areas/suburbs and the neighbourhoods carry the “neighbourhood” names that everyone knows, but the wards are the most significant political unit—at least these three levels should be mapped. There are also large and small-scale “neighbourhoods” that everyone knows by name, but have no official boundaries.

Which of these would be suburbs, and which neighbourhoods? Michael Z. 2011-04-20 19:22 z

We may indeed be able to simply use suburb, as at least the wiki definition seems like it would allow for including neighborhoods. Of course it might be confusing to some to have a place=suburb located inside a place=city, as at least my definition of a "suburb" is an area on the outskirts of a city. -- Joshdoe 04:36, 11 May 2011 (BST)
Reading further into the suburb Wikipedia article, it appears in Ireland and the UK "suburb merely refers to a residential areas outside the city centre", and in Australia and New Zealand "suburbs have become formalised as geographic subdivisions of a city", and in the US and Canada "suburb can refer either to an outlying residential area of a city or town or to a separate municipality, borough, or unincorporated area outside a town or city". This leads me to believe that place=suburb was an unfortunate choice (at least according to the current wiki definition), as it doesn't align with any of these definitions. -- Joshdoe 05:00, 11 May 2011 (BST)
Suburb is a good general-purpose class. It's useful for recognized urban areas in and around cities, regardless of the scale or type of political organization, whether autonomous or subordinate. It's already being used for hierarchies of neighbourhoods, so maybe this is to be embraced. Michael Z. 2011-06-11 16:20 z
While I agree that suburb was a bad decision (it implies outside the 'real' city), it still happens to be well introduced in OSM (currently 52927 uses) so I'd stick with it. I do not agree that we can use this for all kind and size of settlement fractions below city/town. While a suburb might stand on its own, we need something for named parts of quarters/suburbs/districts. My research on this led me to the term neighbourhood, and I was happy to find this page and to see that there is already kind of consensus for this. --Dieterdreist 15:16, 30 August 2011 (BST)

Other keys

Instead of imposing the arbitrary distinction of suburb vs neighbourhood, why not use Key:border_type, or an equivalent, to capture all of the local types of city subdivisions. The complex set of subdivisions I described immediately above could be represented with something like this:

  • place=suburb, admin_level=9, border_type=community area, name=Fort Garry
  • place=suburb, admin_level=10, border_type=ward, name=River Heights–Fort Garry
    • place=suburb, admin_level=11, border_type=neighbourhood, name=Wildwood

(In this real example, the areas of Fort Garry and River Heights–Fort Garry overlap, and Wildwood is within both of them.)

How about an alternative to Tag:boundary=administrative for non-governmental boundaries, like folk-wisdom neighbourhoods or developers' subdivisions? We already have Tag:boundary=civil (but there is no documentation). E.g.:

  • place=suburb, boundary=neighbourhood
  • place=suburb, boundary=subdivision

 Michael Z. 2011-04-20 19:42 z

I don't think that boundary=neighbourhood/subdivision is a good idea. This is not the same kind of information as 'administrative'. If you want to do it with boundary, I'd go for boundary=settlement / settlement-fraction. Still this is not logical. Why should we start subtagging with suburb instead of continuing the place scheme? Place is the key used for settlements, and applying a misleading value like "suburb" and in the next iteration specify that it isn't a suburb but a neighbourhood will not make mapping easier. Why impose 2 keys if one can do it? --Dieterdreist 15:22, 30 August 2011 (BST)

Relation type Site and Neighbourhood mapping (United States)

In the past, I had tried mapping neighbourhoods / subdivisions by placing boundaries. However, the boundaries of subdivisions can be open to dispute and the boundaries lead to significant cluttering of map views. I've recently taken to using a relation to map a neighbourhood / subdivision. My primary source of information is a local government land parcel database which includes subdivision boundaries. I create a relation with the parameters:

As members, I add the streets included ... often only a segment of a street ... without a specified role. If there is a pre-existing place node, usually from USGS GNIS, this gets added to the relation with role=label; if there is not a pre-existing place node, I create one for the label. Also, many subdivisions have only one or two entrances navigable by car, which nodes are included in the relation using role=entrance.


--Ceyockey 11:29, 28 July 2012 (BST)

This is similar to what I've been doing for Winnipeg, Manitoba. I'll have to start including the source information, as you have. I plan to deepen this scheme with the suburb/quarter/neighbourhood hierarchy that's available now.
I've also tagged industrial neighbourhoods with zero population as place=locality, and omitted the label. This results in an appropriately smaller, automatically-generated label. Michael Z. 2012-08-13 15:35 z