Presenter
Sofia Zapounidou (@szapoun)
Abstract
The Publications Office of the European Union (OP) delivers EU Law, Publications and Datasets as a knowledge graph via its open-access repository, Cellar. Despite the OP linked data ecosystem, EU Publications (monographs and serials in any of the 24 EU languages) were catalogued using MARC21, bringing complexity and extra transformations to the dataflow. To address this issue and enrich Cellar with more structured metadata, the ‘Cataloguing Workflow and Tool’ project was initiated. Maintaining disruption-free metadata delivery across all 24 EU languages was the key challenge.
The preparation phase (2021-2024) focused on semantic constructs and data remediation. Although the OP Common Data Model (CDM) is based on FRBR, it initially focused on describing EU Law, lacking constructs for EU Publications. After analysing the semantics of MARC21 Bibliographic elements for monographs and serials, the CDM was updated with the addition of 57 classes and 194 properties. In parallel, to sustain multilingual metadata delivery, a working group replaced local translation tables with semantic EU Vocabularies.
Implementation began in October 2024 selecting the Atom4 software, developed by the Czech company Aion. The OP streamlined the workflow, while Aion customised Atom. Launched in November 2025, the tool is used not only as a linked data editor, but also to share the workload, register DOIs and ISSNs, monitor metadata imports/exports, and generate reports.
This project provides practical insights for migrating from MARC to native linked data cataloguing. The presentation concludes with lessons learned during the first year of operating the new tool, as well as experience gained in multilingual cataloguing using controlled vocabularies and linked data natively.