Lists of prominent and thematic tags?

OpenStreetMap is wonderful for its generic attribute system, though new users such as myself seem to find a lot of difficulty targeting precisely what information is available within the system: is there a set of documentation centered around its applications?

For instance, suppose a user is building a photo-sharing app and wants to populate their photograph metadata with usual conversational text strings associated with the location derived from its lat, lng. The list of potential attributes becomes expansive very quickly, and I do not see many ways to create a comprehensive list. Let me explain.

For instance, a photo-sharing app may benefit from (off the top of my head):

  1. administrative boundaries (names of counties, etc).
  2. National Parks, local topological parks.
  3. bodies of water
  4. theme parks and recreational activities
  5. buildings and chain store names
  6. contextual information to activity (i.e., food inside airport)

Each of these require somewhat radically different attributes to filter their is_in. In some cases, attributes are generic (boundary=national_park), in others they are specific (building=<type>) and others are subcategories of other attributes (amenities=fire_station).

OpenStreetMap is a wonderful community going back more than a decade, so presumably there are resources that elder members of the community have written directly for leading the direction of new users engaging with the community to build applications. Web searches were unable to find any variant of Wiki pages or Medium articles, etc. Where to direct people?

This is a comprehensive list of nearly all tags, much too comprehensive:

Hello David Bernat,

I think that what you are referring here is called a reverse geocoder.

We have Nominatim that can get places based on lat and lng.

However what you are describing here would be a bit more sophisticated.

You can check out taginfo for a more comprehensive list:

I hope that this helps you a bit.

Thank you @Tjuro, those are inputs, yes; though, my question is not asking about Reverse Geocoding (which could be adequately solved by Nominatim).

With a lat,lng the is_in query returns upwards of dozens or hundreds of useful tags. These tags are useful in context specific ways. For instance, knowing that the point is on a bike lane may or may not be useful. Similarly, the convention for extracting tags is extraordinarily tag specific to keep the database general.

Let us talk about an application: providing information to the user about the “place” or “placeness” of the photo. You can see that creating a map from application to useful tags is very hardcoded. “Building type” and “Building function” are accessed using two different tags. Whether inside a “Natural Park” is an entirely different tag, which an entirely different key. The “Body of Water” is actually a filter on two tags in unison. Does this make sense?

Somebody almost certainly has worked through the “application” level discussion of this before: “If you want all the useful imagery content within a building, these are the twenty tag combinations you want,” e.t.c. “If you want to provide geographical region context for a photo sharing app, these are the ten tag combinations you want.” Does my query make sense?

Both the link you shared and I shared point to the same “too comprehensive” information.

Not really a tip of where the secret sauce is shared, here some talk about data-mining openstreetmap - What you map is not always what you get

I think in my business we call that a ‘Wow. OK,’ and disregard that I was doing this for the community. The secret sauce is so straightforward. Thank you for your time. Our answer on our side is trivial.