SPHN Metadata Catalog

Overview

The SPHN Metadata Catalog is a FAIR Data Point-compliant infrastructure that exposes rich, structured metadata in both human and machine-readable form. Dataset descriptions follow the HealthDCAT-AP specification, ensuring standardized representation of health data catalogs across Swiss university hospitals and research institutions. It also includes detailed dataset-level statistics expressed via the Vocabulary of Interlinked Datasets (VoID), giving researchers quantitative insight into dataset composition before requesting access.

Before requesting data access, users can explore the semantic structure of graph schemas and dataset statistics directly through the SPHN Metadata Catalog. Researchers are able to examine data structures, relationships, attributes, and the relative abundances of individual data elements, making it possible to assess feasibility and evaluate dataset suitability and reuse conditions in an informed way.

Approach

• Built on the FDP reference implementation: Used the open-source FAIR Data Point server and FDP client as the foundation, giving us a well-tested, spec-compliant base to build from rather than starting from scratch.
• Use-case specific modifications: Extended and adapted both the server and client to meet SPHN-specific requirements - custom metadata shapes, integration with the SPHN Metadata Catalog Schema, and tailored UI to express additional metadata for researchers.
• Standards alignment: Aligned metadata elements with DCAT, DCAT-AP, and HealthDCAT-AP, ensuring compatibility with Swiss and European health data initiatives and enabling federated catalog discovery across institutions.
• VoID statistics integration: Adapted the underlying representation and UI to display dataset-level statistics (class counts, property distributions, entity abundances) as machine-readable VoID descriptions, giving researchers a quantitative view on dataset structure without requiring data access.
• Stakeholder co-design: Worked directly with data stewards and data managers to validate metadata representations, iterating on the schema to accurately capture relevant information for dataset description.

Tech Stack

• Semantic Web: RDF, RDFS, SHACL
• Vocabularies: DCAT, DCAT-AP, DCAT-AP CH 2.0, HealthDCAT-AP, VoID
• Triplestore: GraphDB, Apache Jena Fuseki
• ETL: Python, RDFLib
• Backend: Java
• Frontend: Vue

Curious to learn more about the SPHN Metadata Catalog? Read about it on our preprint: https://preprints.jmir.org/preprint/90146