Where do geocoding API places get their data from?

Question

A friend of mine asked me to make an application, where he has 12 million addresses, and he needs to look up the closest addresses.

I thought it was trivial, and it really is, except for getting the 12 million address into postgis.

So how do all of these other places know the GPS for an address?

Answer 1

It will be different depending on where you are in the world. In the UK, where I have most experience, addresses are maintained by the Post Office and the actual coordinates by the national mapping agency, which are either surveyed by GPS or inferred from geo-referenced aerial photography.

In the US, people tend to use Tiger and there is an interesting article here about how to load and do geocoding, technically reverse geocoding, with this data; http://www.bostongis.com/blog/index.php?/archives/206-Waiting-for-PostGIS-2.1-Install-PostGIS-Tiger-Geocoder-as-an-Extension.html.

In Western Europe, there is a cadastral system in use, that delineates property boundaries, but I can't tell you much about individual countries and how they actually derive the coordinates.

The Open Street Map project has an experimental project for reverse geocoding based on named streets, which have been collected by GPS in the first place: http://wiki.openstreetmap.org/wiki/Nominatim OpenStreetMap is an amazing collective undertaking, but there are many parts of the world, where it is incomplete.

Answer 2

There is a number of ways companies obtain GPS coordinates for address and such. It is rather complex process of remote sensing in combination with accurate vector mapping.

For general information on the geocoding check out the wikipedia page about it http://en.wikipedia.org/wiki/Geocoding