This repository contains source code and datasets for transparent and reproducible semantic data enrichment of Australian place names data. It is developed by the Geographic Knowledge Lab (GKL). The approach uses RML (RDF mapping language) to create a national placenames knowledge graph based on the Geoscience Australia Placenames Ontology from the publicly available state-based placenames tabular (csv) data.
- Geoscience Australia Place Names Ontology;
- Geoscience Australia Place-Names GitHub repository;
- Composite Gazetteer of Australia;
- Data Product Specification for the Composite Gazetteer of Australia;
- Linked Data API codebase for National Composite Gazetteer of Australia; and
- RML tools.
- data: Folder with data from official gazetteers and place names.
- doc: Project documentation and examples.
- lib: RML processors and dependencies.
- src: RML mapping scripts and data processing codes.
- Gazetteers directly available in the repository for ACT, NSW, NT, QLD, TAS, VIC, and WA.
- For SA only external link available to download official place names gazatteer due to large file size.
- Data downloaded from authoritative organisations (state) for NSW, QLD, SA, VIC, and WA.
- For ACT, NT, and TAS place names gazetteers were downloaded from the national database, the Composite Gazetteer of Australia.
- The list of authoritative and non-authoritative organisations for place name gazetteers is available on the Intergovernmental Committee on Surveying and Mapping (ICSM) website.
| Jursdiction | Metadata | Download Link | Data in GitHub | Last Updated |
|---|---|---|---|---|
| ACT | Web Link | ACT place names in the Composite Gazetteer of Australia | ACT.csv | 25/03/25 |
| NSW | Web Link | Official NSW place names gazetteer | NSW.csv | 13/10/25 |
| NT | Web Link | NT place names in the Composite Gazetteer of Australia | NT.csv | 25/03/25 |
| QLD | Web Link | Official QLD place names gazetteer | QLD.csv | 25/03/25 |
| SA | Web Link | Official SA place names gazertteer | Sites.csv; Lines.csv; Areas.csv | 25/03/25 |
| TAS | Web Link | TAS place names in the Composite Gazetteer of Australia | TAS.csv | 25/03/25 |
| VIC | Web Link | Official VIC place names gazetteer | Sites.csv; Roads.csv | 25/03/25 |
| WA | Web Link | Official WA place names gazetteer | WA.csv | 25/03/25 |
The below image shows the snapshot of the classes, Object Property (OP), and Data Property (DP) of the Geoscience Australia Place Name ontology. Defined relations in the ontology are used for RML mapping and building PNKG. In the figure below, yellow circles represent classes, blue rectangles indicate object properties, and green rectangles depict data properties.
The knowledge construction workflow includes the following stages:
- data preparation;
- RML mapping and processing; and
- knowledge graph generation.
The following figure illustrates the workflow for generating the knowledge graph.
This folder contains a link to download the Australian Place Names Knowledge Graph (PNKG), which integrates gazetteer data for all Australian states and territories. The figure below presents a snapshot graph of a section of the PNKG.
A web application was developed to retrieve and display additional contextual information about Australian place names stored in a Knowledge Graph. The GeoSPARQL Fuseki Server acts as the SPARQL endpoint. When a user clicks on a place on the map, the application displays the enriched contextual information in a separate Knowledge Graph Explorer view. More information about the technical workflow and dependencies can be found in the webInterface subfolder.
The RML mapping and processing stage supports only data in CSV format. To use the same mapping rules, data in other formats must be converted to CSV. Additionally, minor modifications to the mapping rules may be required. Further work under consideration also includes development of SHACL rules and validate the RDF output for consistency and correctness.
This development of pnkg was led by Nayomi Ranamuka and Nenad Radosevic, with contributions from all of the researchers at the RMIT Geographic Knowledge Lab: Alexis Horde Vo, Prawal Lohani, Amina Hossein, Mohammad Kazemi Beydokhti, Ozzy Yaguang Tao, and GKL Director Prof Matt Duckham. We're also grateful to the 2024 OGC Metadata Code Sprint, and to advice from Nicholas Car and Rob Atkinson.