I'm looking for a dataset that contains a complete hierarchical geographic representation of the world. Basically, I want names, boundaries, relationships ("is a member of"), ISO codes, and more for:

  • continent (optional but useful)
  • country
  • state/province/territory
  • district/county
  • city

Postal code would be nice, too, but that can complicate things since postal codes can sometimes span multiple regions (e.g., ZIP code 02861 covers area in both Pawtucket, Rhode Island and Attleboro, Massachusetts).

Some sort of globally unique ID for each location would be really useful. ISO 3166 specifies country codes (3166-1) and sub-country (state/province/territory generally) codes (3166-2), but it doesn't seem to contain any finer administrative units. Is there an ISO standard for finer administrative areas than state/province/territory? For example, is there a standard (ISO?) that labels the following administrative areas?:

  • Orange County, CA, USA
  • Anaheim, CA, USA (a city within Orange County)
  • Bengbu District, Anhui Province, China

Basically, I want a full geographic hierarchy of the entire world. Does such a dataset exist?

Things I've explored:

  • Natural Earth - Great open source dataset that contains good data on countries, states, and cities. It doesn't offer county/district data, though, and there doesn't appear to be any sort of standard global location ID. It also doesn't offer boundary data for cities; instead, cities are represented by a point (centroid?).
  • Universal Postal Union POST*CODE Database - From what I can tell, this has global postal code data, but I'm not sure where/how I can get the full dataset; all I see are a few sample files.
  • UN/LOCODE - This is a really extensive dataset that seems to have lots of cities and special locations (Air Force bases, airports, etc.) and even offers a unique 3-letter identifier for each, but it doesn't seem to have counties.

What else should I be looking at? Surely such a dataset exists, right?

  • 2
    Perhaps this question would be a better fit in the Open Data Stack Exchange. – Kingfisher Feb 25 '16 at 22:46
  • Wow, I didn't even know opendata.stackexchange.com existed! I'll wait and see if anyone here has any ideas before cross-posting. Thanks for the suggestion. – Geoff Feb 25 '16 at 22:49
  • Have you tried geoName.org? – Ian Turton Feb 27 '16 at 15:59
  • @iant, I sure haven't. This looks incredible useful, actually! – Geoff Feb 29 '16 at 18:02

This is a tough christmas wish!

Read this https://mapzen.com/documentation/search/data-sources/ and this blog: https://mapzen.com/blog/who-s-on-first to learn about OpenStreetMap sources and the (few) additional sources.

Then look especially at the very recent Zetashapes: http://zetashapes.com/

Bottom-line is: OpenStreetMap will be the dataset you want, but it's neither complete nor easy to access at this time.

| improve this answer | |
  • You've provided some really useful links through that Mapzen post. GeoNames looks especially useful for me. It looks like they allow people to buy their data here. Quattroshapes also looks really great. I think you've answered my question, so I'm gonna go ahead and mark this as the answer. Thanks! – Geoff Feb 29 '16 at 18:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.