GeoResolver is a lightweight Python library for resolving place names into geographic coordinates and related metadata using multiple gazetteer services, including GeoNames, WHG, Wikidata, and TGN.
The library provides a unified interface and standardized response format across sources, making it easier to disambiguate, enrich, or geocode place names—especially in datasets, archival collections, and manually curated records.
GeoResolver is particularly useful for historical geocoding and legacy datasets, where place names may be inconsistent, ambiguous, or obsolete. It is not intended to replace tools like Geopy for general-purpose geocoding. Instead, GeoResolver complements them by offering a targeted approach based on authoritative gazetteers, tailored for academic, historical, and archival contexts.
The logic behind GeoResolver is straightforward:
Given a place name as input, the library queries one or more gazetteers in sequence, searching for the closest match using a fuzzy matching algorithm. If a sufficiently good match is found, it returns the coordinates of the place. If not, it moves on to the next gazetteer, continuing until a match is found or all gazetteers have been queried.
If no match is found in any gazetteer, the library returns a None value.
A fuzziness threshold can be configured to control how strict the match should be. The default threshold is 90, meaning the library only accepts matches that are at least 90% similar to the input. Lowering the threshold allows more lenient matches; raising it makes the match stricter.
It's possible to be even more flexible by enabling a flexible threshold for short place names. This is useful when you want to allow some places (like 'Rome', or 'Lima') to match without reducing the threshold for longer names.
To improve precision, you can filter by country code and place type for deambiguation or to narrow down results.
Some services allow specifying place types using localized terms, which can be useful when working with multilingual datasets.
GeoResolver includes a basic mapping of common place types in data/mappings/places_map.json. You can also pass a custom mapping to the PlaceResolver class to support additional types or override defaults. This is useful for adapting the resolution logic to domain-specific vocabularies or legacy data.
To use GeoResolver, install the library via pip. It’s recommended to use a virtual environment to avoid conflicts with other packages:
pip install georesolverTo use the GeoNames service, you must create a free account at GeoNames and obtain a username. This username is required to make API requests.
Warning: It's possible to use the username
demofor testing purposes, but this user has very limited quota and it's possible to hit the limit quickly, especially with batch requests.
You can provide your username in one of two ways:
Environment variable
Create a .env file in your project directory:
GEONAMES_USERNAME=your_geonames_username
Pass it explicitly
from georesolver import GeoNamesQuery
geonames_query = GeoNamesQuery(geonames_username="your_geonames_username")The most straightforward way to use the library is through the PlaceResolver class. By default, PlaceResolver queries all available services — GeoNames, WHG, TGN, and Wikidata — in that order.
To resolve a place name, call the .resolve() method with the name and (optionally) a country code and place type. If no filters are specified, the first sufficiently similar match across all services is returned.
from georesolver import PlaceResolver
# Initialize the resolver (uses all services by default)
resolver = PlaceResolver()
# Resolve a place name
result = resolver.resolve("London", country_code="GB", place_type="inhabited places")
if result:
print(f"Coordinates: {result['latitude']}, {result['longitude']}")
print(f"Source: {result['source']}")
print(f"Confidence: {result['confidence']}")
else:
print("No match found")Sample output:
Coordinates: 51.50853, -0.12574
Source: WHG
Confidence: 100.0Starting with v0.2.0, the resolve() method returns a structured dictionary with comprehensive metadata:
{
"place": "London",
"standardize_label": "London",
"language": "en",
"latitude": 51.50853,
"longitude": -0.12574,
"source": "GeoNames",
"id": 2643743,
"uri": "http://sws.geonames.org/2643743/",
"country_code": "GB",
"part_of": "",
"part_of_uri": "",
"confidence": 95.5,
"threshold": 90,
"match_type": "exact"
}You can control which services PlaceResolver uses and configure them individually. For example:
from georesolver import PlaceResolver, GeoNamesQuery, TGNQuery
geonames = GeoNamesQuery(geonames_username="your_geonames_username")
tgn = TGNQuery()
resolver = PlaceResolver(
services=[geonames, tgn],
threshold=80,
flexible_threshold=True, # Use flexible threshold for short place names
flexible_threshold_value=70, # Lower threshold for short names
lang="es", # Spanish language support
verbose=True
)This gives you more control over the resolution logic, including match strictness (threshold), flexible thresholding for short place names, language preferences, and logging verbosity (verbose=True).
GeoResolver supports batch resolution from a pandas.DataFrame, making it easy to process large datasets.
You can use the resolve_batch method to apply place name resolution to each row of a DataFrame. This method supports optional columns for country code and place type, and can return results in different formats.
import pandas as pd
from georesolver import PlaceResolver, GeoNamesQuery
# Sample data
df = pd.DataFrame({
"place_name": ["London", "Madrid", "Rome"],
"country_code": ["GB", "ES", "IT"],
"place_type": ["city", "city", "city"]
})
# Initialize the resolver
resolver = PlaceResolver(services=[GeoNamesQuery(geonames_username="your_username")], verbose=True)
# Resolve in batch, return structured results as a DataFrame
result_df = resolver.resolve_batch(df,
place_column="place_name",
country_column="country_code",
place_type_column="place_type",
show_progress=False
)
print(result_df.columns.tolist())
# Output: ['place', 'standardize_label', 'language', 'latitude', 'longitude', 'source', 'place_id', 'place_uri', 'country_code', 'part_of', 'part_of_uri', 'confidence', 'threshold', 'match_type']This returns a new DataFrame with columns for all resolved place attributes including coordinates, source information, and confidence scores.
The resolve_batch method returns a pandas.DataFrame by default, but you can also return a list of dictionaries that can be useful for JSON serialization.
# Return results as a list of dictionaries
results = resolver.resolve_batch(df,
place_column="place_name",
country_column="country_code",
place_type_column="place_type",
return_df=False, # Return list of dictionaries
show_progress=False)
print(results[:2])Example output:
[
{
"place": "London",
"standardize_label": "London",
"language": "en",
"latitude": 51.50853,
"longitude": -0.12574,
"source": "GeoNames",
"id": 2643743,
"uri": "http://sws.geonames.org/2643743/",
"country_code": "GB",
"part_of": "",
"part_of_uri": "",
"confidence": 100.0,
"threshold": 90,
"match_type": "exact"
},
{
"place": "Madrid",
"standardize_label": "Madrid",
"language": "en",
"latitude": 40.4165,
"longitude": -3.70256,
"source": "GeoNames",
"id": 3117735,
"uri": "http://sws.geonames.org/3117735/",
"country_code": "ES",
"part_of": "",
"part_of_uri": "",
"confidence": 100.0,
"threshold": 90,
"match_type": "exact"
}
]Different gazetteers use different terms to classify place types (e.g., "populated place", "settlement", "city", "pueblo"). To unify these differences, GeoResolver uses a configurable place type mapping that standardizes input values before querying services.
By default, GeoResolver uses a built-in mapping stored at data/mappings/places_map.json. This file maps normalized place types (like "city") to the equivalent terms used by each service.
Example mapping entry:
"city": {
"geonames": "PPL",
"wikidata": "Q515",
"tgn": "cities",
"whg": "p"
},You can provide your own mapping by passing a JSON file path to PlaceResolver:
resolver = PlaceResolver(
services=[GeoNamesQuery(geonames_username="your_username")],
places_map_json="path/to/your_custom_mapping.json"
)This is useful when working with domain-specific vocabularies, legacy datasets, or non-English place type terms. You can also use it simply to override the default mapping with your own preferences.
Each service-specific list should contain valid place type codes or labels expected by that gazetteer.
This library queries the Wikidata MediaWiki API via the endpoint:
https://www.wikidata.org/w/api.php
It does not use the SPARQL endpoint (https://query.wikidata.org/sparql), as this approach is faster and more reliable for simple place lookups. The library performs entity searches by name and retrieves coordinates, country (P17), and administrative data from the entity information.
Enhanced in v0.2.0: WikidataQuery now provides better country and administrative entity data retrieval, with improved matching against the BaseQuery interface for consistency across all services.
⚠️ Performance Note: Wikidata API queries involve multiple HTTP requests per place (search + entity data). This process is relatively slow and not recommended for bulk resolution. Consider using GeoNames or WHG for large-scale batch processing.
Contributions are welcome! If you encounter a bug, need additional functionality, or have suggestions for improvement, feel free to open an issue or submit a pull request.
This project is licensed under a GNU General Public License v3.0. See the LICENSE file for details.
This library relies on open data sources and public APIs provided by GeoNames, WHG, Wikidata, and TGN. Special thanks to the maintainers of these projects for their commitment to accessible geographic knowledge.