Case study: INFRA-ART metadata crosswalks #110
Description
The title of your case study
Semantic Artefact Documentation for the INFRA-ART Spectral Library Datasets
Short title
INFRA-ART metadata crosswalks
Summary
This case study presents the development of a comprehensive set of semantic artefacts—Schema.org/DCAT JSON-LD schemas and ontology mappings—designed to enhance the machine-actionability and semantic interoperability of the existing datasets within the INFRA-ART Spectral Library, a FAIR-oriented database that supports heritage research specialists working with spectroscopic techniques. The work aligns local dataset descriptors with the EOSC EDMI metadata set and integrates more than 50 domain-specific terms mapped to established ontologies such as CHMO, OM, AAT, EDAM, and others.
Domain
Heritage Science / Interdisciplinary
Use case category
Semantic interoperability (shared understanding of data across multiple systems)
Purpose of the mapping
The mapping was undertaken to address the lack of semantic interoperability in the existing datasets of the INFRA-ART Spectral Library. Its primary purpose was to improve machine-actionability and discoverability by ensuring that each spectral collection is described using structured, machine-readable metadata that supports automated access, interpretation, and reuse. An additional aim was to align the repository with EOSC interoperability requirements, enabling seamless integration with EOSC-aligned catalogues and aggregators.
Type of mapped resources
The core mapping activity involves aligning the INFRA-ART dataset descriptors with the EOSC EDMI metadata properties and expressing them through DCAT and Schema.org for machine-actionable JSON-LD exposure. In parallel, more than 50 domain-specific descriptors are mapped to ontology terms from established vocabularies such as CHMO, OM, AAT, EDAM, and FIX, ensuring semantic consistency and cross-domain interpretability.
Links to an existing mappings
https://zenodo.org/records/17570835
Tools used for creating the mapping
The mapping was created through manual curation using Excel, with ontology lookup supported by the EMBL-EBI Ontology Lookup Service and the Getty Vocabularies. JSON-LD drafts were prepared and refined with support from LLM-based assistants. Metadata validation was performed using the Google Rich Results Test, the Schema.org Validator, and F-UJI. Alignment of EOSC EDMI properties with Schema.org and DCAT terminology was guided by the RDA Research Metadata Schema Crosswalk Mappings to Schema.org (https://fairsharing.org/3641).
Type of mapping relations
The mappings consist mainly of schema-level equivalence relations between the internal descriptors of INFRA-ART’s spectral datasets and their corresponding properties in EDMI, Schema.org, and DCAT, using primarily one-to-one and one-to-many correspondences. Ontology mappings apply semantic relations such as skos:exactMatch, skos:closeMatch, and skos:broadMatch to align dataset descriptors with external vocabularies (CHMO, OM, AAT, EDAM, FIX). No unit conversions or complex transformation functions were required.
Examples (samples) of different types of mapping implementations
infraart:instrument → edm:instrumentType → measurementTechnique
infraart:material → edm:objectType → dct:subject (AAT URI)
infraart:wavelengthRange → edm:variableMeasured → variableMeasured
infraart:datasetURL → edm:landingPage → dcat:landingPage