A Fair Pipeline from Archival APIs to Person Reconstructions: Linked Data, Transparent Linkage, and Queryable Provenance

Richard L. Zijdeman, International Institute of Social History
Rick J. Mourits, International Institute of Social History
Xander Wilcke, Vrije Universiteit Amsterdam
Mateusz Kielan, International Institute of Social History

Growing awareness of FAIR principles has reshaped expectations for reproducibility and reuse in research, yet practical implementations tend to concentrate on either the research application layer, such as executable notebooks, or on dataset description and publication. We identify two critical gaps: archival sources themselves must be integrated into FAIR workflows, and FAIR practices have not been systematically applied to person reconstruction, a domain characterized by heterogeneous methods and substantial opportunity for cross-study learning. This paper demonstrates a complete FAIR-aligned pipeline that spans from archival API to published person and family reconstructions. Starting from archival APIs we extract person observations from records and represent them as linked data using PiCo, a community-designed vocabulary that leverages Schema.org concepts to maximize interoperability and reusability. We then apply burgerLinker, an open-source linkage tool that produces person reconstructions while recording the generic linkage choices used, including blocking strategies and relative string-distance measures, enabling replication and quantitative comparison of linkage configurations. Data quality is enforced and monitored with SHACL constraints, and burgerLinker produces structured metadata to support efficient batch review of reconstructions. Final outputs are exposed through a SPARQL endpoint to enable community queries on persons and families; the integrated pipeline also provides direct provenance to archival scans so that anomalies can be resolved with a single click back to the original source. Together, these contributions show how embedding archives, transparent linkage, validation, and queryable linked data into a unified workflow advances FAIR practice for person reconstruction and fosters cumulative, comparable research across the field.

No extended abstract or paper available

 Presented in Session 144. Beyond the Census II