How to Geocode Australian Addresses with GeoPy & Open Land Fabric

Learn how to geocode Australian addresses using GeoPy's ArcGIS provider with Open Land Fabric. Covers setup, Pandas batch processing, and concurrency best practices.

Open Land Fabric Admin

— 1 min read

GeoPy is the most popular Python library for geocoding, but standard commercial providers require credit cards, paid API tokens, or restrictive attribution terms. Open Land Fabric provides a free, unkeyed, native ArcGIS-compatible GeocodeServer surface covering all 16.1M Australian addresses and 8.6M cadastre parcels.

Zero Setup Required: Because Open Land Fabric implements standard Esri REST specifications, simply pass domain="openlandfabric.org" to GeoPy's built-in ArcGIS geocoder.

1. Quickstart: Geocoding in Python

Install GeoPy via pip or uv:

pip install geopy

Then run a forward geocode query:

from geopy.geocoders import ArcGIS

geolocator = ArcGIS(domain="openlandfabric.org", timeout=10)
location = geolocator.geocode("100 Flinders Street, Melbourne VIC 3000")
print(location.address)
print(f"Latitude: {location.latitude}, Longitude: {location.longitude}")

2. Geocoding Pandas DataFrames with RateLimiter

When geocoding lists in Pandas, wrap the geocoder in GeoPy's RateLimiter to pace requests sequentially and retry transient network errors:

from geopy.extra.rate_limiter import RateLimiter

geocode = RateLimiter(geolocator.geocode, min_delay_seconds=0.15, max_retries=2)
df["location"] = df["address"].apply(geocode)

3. Concurrency Protections vs. Batch Processing

In-flight requests are capped at 4 concurrent connections per caller to prevent thread exhaustion. For 500+ rows, use the Client Workspace or Batch Geocoding Service.

4. Reverse Geocoding

location = geolocator.reverse("-37.8164, 144.9702")
print(location.address)

Read more