GNIS-LD was created by the UCSB STKO Lab. This effort led to the creation of a number of software products, listed in the source code section. Among these included the full GNIS-LD ETL pipeline with a web interface. This team was composed of

Blake Regalia
Krzysztof Janowicz
Gengchen Mai
Dalia Varanka
E. Lynn Usery

The project is currently maintained by DataONE through funding from NSF grant 2033521 under the KnowWhereGraph project. With this change, the infrastructure was extended to deploy on Kubernetes and the website and various software components were updated.


The original project published two papers about work in this area. These are recommended reading to learn more about project motivations and the original software stack.

Regalia, B., Janowicz, K. , and McKenzie, G. (2019): Computing and Querying Strict, Approximate, and Metrically-Refined Topological Relations in Linked Geographic Data. Transactions in GIS.

Regalia, B., Janowicz, K. , and Mai, G. (2018): GNIS-LD: Serving and Visualizing the Geographic Names Information System Gazetteer As Linked Data. 15th ESWC Conference. 3-7 June, Heraklion, Crete.

Current Source code

The source code for the current GNIS-LD project is hosted on GitHub by DataONE.

The Kubernetes deployment responsible for container orchestration and networking.
Converts GNIS into RDF.
This website which includes dereferencing logic, redirection, and SPARQL query proxying.

Original Source Code

The source code for the GNIS-LD project, developed at the UCSB STKO Lab can be found below.

Bundles the entire pipeline into a Docker Compose file and a helper bash script. Its components are listed out below.
Converts GNIS and other USGS datasets (such as NHD) into RDF.
Loads triples directly into KiWi's PostgreSQL tables in parallel.
The 584 branch of Apache Marmotta supports GeoSPARQL.
This website which includes dereferencing logic, redirection, and SPARQL query proxying.


For any questions about the service, contact DataONE Support (