Where is Wikipedia's coordinate information stored, and how can I retrieve it?

Question

Background: I'd like to add icons to my map showing locations which have Wikipedia entries.

The WikiLocation API is a great resource for determining Wiki entries near an XY location, but it has limitations (20km radius, 50 points only) which make it unsuitable for my purposes.

Ideally I need the map to show ALL locations with Wiki entries, so I guess I'll need to download the Wikipedia dump file and parse it for geocoded entries.

My question: where is this information found in the Wikipedia dump files?

The Wiki database may be downloaded from http://en.wikipedia.org/wiki/Wikipedia:Database_download but I can't see which file(s) I need to search for the coordinates.

For example, the entry for Sydney includes:

Coordinates:    33°51′35.9″S 151°12′40″E

Where is this found in the dump files? Another way to phrase the question is "how can I determine which Wiki entries are geo-coded?"

Thanks.

Answer 1

See the database dump section on http://de.wikipedia.org/wiki/Wikipedia:WikiProjekt_Georeferenzierung/Wikipedia-World/en. This gives the option for both a CSV download and a postgreSQL connection.

Also look through http://en.wikipedia.org/wiki/Wikipedia:WikiProject_Geographical_coordinates for other options.

Answer 2

An API solution is probably your best bet, but if you are dead-set on using the database dumps, in this Twitter response the WikiLocation API author explains a bit that he is not scraping the entire Wikipedia database dump but only certain tables.

I would start by looking at the external links database and looking for the links to GeoHack (probably only ones matching the pattern used by such URLs as the Sydney one).