Python geocoding in a disagreeing world
Geocoding is the process of taking a string representing a real-world address, and resolving that to a single point location. However different providers will resolve the same addresses to different locations, and it can be important to consider this spatial error. For example, when determining the census tract an addresses should be linked to, there may be some error. This error can go unacknowledged but still be very influential on analysis, similar to the famous modifiable area-unit problem (MAUP), but for point representations of addresses.
This library can be used to determine spatial error across selected geocoding providers. It can, for example, produce the convex hull of successful geocodes, so you can imagine an address as a polygon. These individual points are also available as a multipoint geometry.
Intended to work with as many of the providers supported by geopy as you care to configure. Without configuration, will use free global provdiders that don’t require API tokens.
Only supports Python 3. Tested with Python 3.4 and 3.5.