Our library to maintain our locations data files, data format, and search interface. We currently have data from Geonames and from OpenStreetMaps.
Run:
$ python setup.py install
A simple usage example:
from geonamescache.geonames.data_source import DataSource
data_source = DataSource()
print data_source.all_locations_search('Lebanon')
print data_source.get_location_by_id(6252001) # USA
Data from http://download.geonames.org/export/dump/. Using the cities5000 data set, this contains ~50000 cities, ~40000 admin level 2 districts, ~4000 admin level 1 districts, and ~250 countries. The above link also documents the fields provided for each type of location.
We currently use Geonames as the data source in primer-core.
See the README in geonamescache/geonames for how to manage the data.
Data from https://github.com/OSMNames/OSMNames/releases/tag/v1.1. This contains the 100K most important locations according to counts of wikipedia links, including ~75000 cities, ~15000 admin level 2 districts, ~2000 admin level 1 districts, and ~250 countries. http://osmnames.org/download/ documents the fields provided for each location.
Note: this data source appears to be missing some key locations such as Vienna, Bangkok, Seoul, Cairo, etc. as well as some fields such as the admin level 1 of Chicago. (We think the missing locations is from having an incorrect importance score derived from wikipedia data, and that nominatim has more accurate data).
We also have scripts to
- fetch alternate names for locations from wikipedia
- compute estimated importance scores for locations for the Geonames data set