Batch Geocoding with Python

You have a list of addresses, but you need to get GPS coordinates to crunch some numbers. Don’t despair, there is geocoding for this and Python provides some simple means to help to deal with the APIs out there.

One way to convert addresses is by using one of many geocoding APIs such as Google. This can be done by using the HTTP library requests, which can be used to access the Google Geocoding Service as in the next snippet.

import requests
url = ''
params = {'sensor': 'false', 'address': 'Centre Pompidou, Paris, FR'}
r = requests.get(url, params=params)
results = r.json()['results']
location = results[0]['geometry']['location']
lat, lng = location['lat'], location['lng']

An important thing to note for a large batch of addresses is that there are usage limits for this API (2500 free requests per day and 50 requests per second). If these limits are reached, the API returns OVER_QUERY_LIMIT response.

Another way of retrieving latitude and longitude of addresses is by using the Python Geocoder library which is consistent across multiple geocoding providers and quite easy to use. It also provides reverse geocoding, geocoding of IP addresses and some other handy tools which can be found in the documentation. As for our main goal, batch geocoding with the Google provider can be done in the following way.

import geocoder
import time

addresses = ['Centre Pompidou, Paris, FR', 'Times Square, NY', ...]
# Use a dictionary to collect all the geocoded addresses
coords = {}

for address in addresses:

    # Apply some sleep to ensure to be below 50 requests per second

    g =
    if coords not in address:
        if g.status == 'OK':
            coords[address] = g.latlng
            print('status: {}'.format(g.status))

Important to note is that for some providers and some addresses it sometimes helps to run multiple trials if OVER_QUERY_LIMIT or ZERO_RESULTS comes as a response for g.status. Another thing to note is that some addresses with “city” in the string might return the city centroid coordinates if the address is ambiguous. This can be easily checked if different addresses result in the same GPS coordinates. Due to the limit of the Google provider, I tend to use the ArcGIS Provider which uses the ArcGIS REST API.

The title image comes from a visualization with the data of the Austrian monuments registry (Bundesdenkmalamt Denkmalverzeichnis) where I used geocoding with ArcGIS and I visualized the data using the Python API of Blender. The project can be seen in this repository and the code for geocoding can be found in the same repository in

Image from austrian-monuments-visualization