Proposed features/new place values

From OpenStreetMap Wiki
Jump to: navigation, search
new place values
Status: Abandoned (inactive)
Proposed by: Dieterdreist
Tagging: rank=*
Drafted on: 2011-01-18

the current order of settlements in place=*:

The place values for settlements are to be considered relative to the structure of the country, and should therefore be defined by country. In comparison to other sources the current place values form a hierarchy with very few levels. Having just 2 main types of cities is not sufficient to render low zoom maps (and creates undesired problems also in higher zoom levels). Currently the official mapnik rendering overcomes this sad situation with external sources (from natural earth).

Tag Population Description Notes
place=city 100,000+
place=town 10,000 - 100,000 an urban settlement with local importance often the town-status is historically grown and defined by legislation, independent from the population numbers
place=village <10,000 incorporated municipality, regardless of its population.
place=hamlet <100 unincorporated settlement with less than 100 inhabitants
place=isolated_dwelling not more than 2 households the smallest kind of human settlement.

Introducing rank as an additional key to define a finer grained structure

Rank is mostly about hierarchy, not in all cases the population numbers will reflect this. You should adopt the scheme and use the numbers as orientation but not as hard limits. Tag based on the context and have a look how other mappers in your region have judged. Use the mailing lists and forum to discuss in case of doubt.

Ideally the community of each country will judge on the classification of their settlements, maybe lists could be made in the wiki, at least for the cities. The top level (place=city and rank=0) should be internationally agreed on. There are different aspects that should be integrated in the evaluation (economy, infrastructure, culture, politics/administration/organizations/NGOs, religion).

General advantages

This approach is consistent with the current tagging scheme. As it is an additional key, it doesn't corrupt the current rendering in the introduction time.

village

The population numbers below are meant to be solely indicative and intended for the situation in Central Europe. Other regions might have different limits. Also there might be agreements to upgrade or downgrade a settlement based on evaluation of the context. The key rank is intended to define the general importance of a place (or other feature) i.e. refine the existing values.

  • village & rank=30 population smaller then 500
  • village & rank=20 population 500 - 2,000
  • village & rank=10 population 2,000 - 5,000)
  • village & rank=0 population > 5,000

town

The population numbers below are meant to be solely indicative and intended for the situation in Central Europe. Other regions might have different limits. Also there might be agreements to upgrade or downgrade a settlement based on evaluation of the context. The key rank is intended to define the general importance of a place (or other feature) i.e. refine the existing values.

  • town & rank=30 population smaller then 10,000
  • town & rank=20 population 10,000 - 20,000
  • town & rank=10 population 20,000 - 50,000
  • town & rank=0 population 50,000 - 100,000

city

  • city & rank=30 population 100,000 - 500,000
  • city & rank=20 city with it's own metropolitan area
  • city & rank=10 population > 1 million in the urban agglomeration
  • city & rank=0 metropolitan area with more than 10 million people, global city, high international importance (economy, transport, culture, politics, religion)

Lists on tagging rank for cities

Research on the ranking of cities

There is lots of studies regarding this topic. Dependent on which criteria they focus they do have different outcome. These are some studies:


These are lists for each country (if yours is missing feel free to add it) to collaboratively agree on an importance level. They are to be considered work in progress.

Europe

Germany

  • place=city & rank=0 Berlin (could get also a slightly inferiour value then London or Paris, say 1)
  • place=city & rank=10 Hamburg, München, (Köln), (Frankfurt) (more or less should be undisputed, not sure for cities in brackets)
  • place=city & rank=20 Nürnberg, Hannover, Düsseldorf, Stuttgart, Dortmund, Essen, Bremen, Leipzig, Dresden (open for discussion)
  • place=city & rank=30 the rest of the German cities

Italy

  • place=city & rank=0 no cities (probably Rome could be rank=0 because it is the world capital of catholicism), Milan could also get in this category because it is quite important economically, for the fashion world and has a big metropolitan area
  • place=city & rank=10 Milano, Roma, Napoli
  • place=city & rank=20 Torino, Firenze, Palermo, Venezia, Genova, Bologna, Bari, Catania (open for discussion)
  • place=city & rank=30 the rest of the Italian cities

feel free to add a list for your country and to discuss the lists that are already present here.

North America

USA

place=city and rank=0

  • New York City
  • Los Angeles
  • Chicago
  • Washington D. C.

place=city and rank=10

  • San Francisco
  • Dallas
  • Philadelphia
  • Boston
  • Detroit
  • Houston
  • Atlanta
  • Miami
  • Phoenix
  • Seattle
  • Minneapolis-Saint Paul
  • Cleveland
  • Denver
  • St. Louis
  • Portland
  • Pittsburgh
  • Las Vegas

place=city and rank=20

  • Tampa, Florida

feel free to add suggestions on the discussion page or to modify here

place=city and rank=30

to be defined by the local mappers

Canada

place=city and rank=0

Nothing. Maybe Toronto? Or Ottawa to see the capital from low zoom levels?

place=city and rank=10

  • Toronto
  • Montreal
  • Vancouver
  • Calgary
  • Ottawa

place=city and rank=20

  • Québec
  • Hamilton
  • Edmonton
  • Winnipeg
  • Regina

Incomplete and not implemented

Other approaches to treat with this problem

This problem has been discussed for years in OSM but few thing have changed in the past 3 years. One of the most given answers is: solve it by processing the data. I do agree that it might be possible to do massive processing and achieve a reasonable result with just a few problems (edge cases that would have to be solved manually in traditional cartography). But this approach is complicated and requires some time and a lot of computational power as well as a fairly complete map data. It is not an alternative to refining the structure, but an additional way one might choose to ultimately solve the problem. One simple approach would be to normalize population=* and order by this, but a lot of cases are not so easy and this would not lead to good results as long as these figures are not compared to their surroundings.

Other pages in this wiki

rank on other features

Could also be used in conjunction with other features, where it is difficult to determine relative ordering automatically from the data, e.g.

  • rivers
  • airports
  • memorials
  • monuments
  • all kind of other features

...

starting with rank=0 at the higher end. Proposed to initiate with multiples of 10 (permitting eventual refining later)

See Also