Dataset of the Multilingual Gazetteer of the Settlements on the Livonian Coast of Northern Kurzeme

This dataset provides a multilingual, semantically linked gazetteer of historical and present-day settlements along the Livonian coast of Northern Kurzeme, Latvia. It supports digital humanities research by aligning toponyms across Livonian, Latvian, German, Russian, Estonian and Finnish sources, and linking each locality to persistent, dereferenceable identifiers. Entries are modelled as RDF resources and enriched with cross-references to external authorities such as GeoNames, VIAF, and Wikidata, ensuring interoperability with global knowledge graphs and cultural heritage infrastructures.

Developed within the Finno-Ugric Data Sharing Space (FUDSS), the gazetteer accounts for historical naming variations and administrative boundary changes over two centuries. It offers a reproducible model for constructing semantically rich gazetteers in minority-language and multi-ethnic regions, with potential applications in endangered language documentation, textile provenance research, and regional heritage studies.

Included files and formats

• livonian_coastal_villages_020.csv — a simplified, tabular view for comparative study of village names in six languages (Livonian, Latvian, Estonian, Finnish, Hungarian, and German). Each record includes persistent identifiers from GeoNames, Wikidata, and the Finno-Ugric Data Sharing Space (FUDSS), providing links to the corresponding structured RDF resources. Accompanied by   livonian_coastal_villages_020_csv-metadata.json describing this serialisation with the CSVW Metadata Vocabulary.

• livonian_coastal_villages__020.ttl and livonian_coastal_villages_020.rdf — the faithful RDF dumps in Turtle and RDF/XML syntax. These preserve Wikibase-internal semantics, including statement nodes, property ranks, and qualifiers, and can be parsed back into a Wikibase instance without loss of information.

• livonian_coastal_villages_truthy_020.nt — a truthy RDF dump using dereferenceable property IRIs (https://reprexbase.eu/fu/Special:EntityData/P###). This format is optimised for loading into generic triplestores (e.g., Blazegraph, Fuseki, GraphDB) and for use in Linked Data pipelines. It omits Wikibase-specific metadata such as statement ranks, internal qualifiers, and references, but remains fully valid RDF and machine-interoperable.