Indexing Historical Research Data: MemO and the NFDI4Memory Knowledge Graph

The NFDI4Memory consortium is part of the NFDI (National Research Data Infrastructure) framework and aims to create long-term and sustainable research data infrastructures for the historically oriented sciences [1, 2]. To bridge the gap between the wide range of disciplines and to connect distributed and heterogeneous sources, the participating task areas strive to build a networked and FAIR digital research environment. Fundamental elements of this endeavour are the Memory Ontology (MemO) and the Knowledge Graph (KG). They form the underlying framework for (federated) search and are brought together in the NFDI4Memory Data Space, 4Memory's digital infrastructure for scientific access to research data and services [3]. The KG acts as an index, linking data from decentralised research sources. It represents research data, institutions, researchers, their expertise, and services offered. MemO, based on NFDIcore and the Culture Ontology (CTO) - both maintained by NFDI4Culture - provides the basic structure [4, 5]. NFDIcore has been developed as a mid-level ontology for representing metadata related to NFDI resources, including individuals, organisations, or projects. The CTO module extends this ontology by adding concepts relevant to the cultural heritage domain [6]. MemO is a modular extension of the above ontologies, adding concepts from the historical sciences. It has been developed based on competence questions collected from the community. MemO's concepts harmonise metadata for resources to be found with a single search, linking fragmented resources based on common concepts such as people, places, or themes, with particular attention to provenance representation. In the currently available version of the KG, central services of the participating institutions are linked, e.g., via common disciplines or technologies such as standards like RDF/XML or common endpoints like OAI-PMH. Suitable services can be found with a single SPARQL query, and further information can be gathered with a federated search. The same idea applies to the research data provided by participating institutions to link fragmented resources through shared information. An example of this is the resources provided by the European Holocaust Research Infrastructure (EHRI) portal and Archivportal-D. EHRI provides information on Holocaust-related archival materials held in institutions around the world. Archivportal-D provides digital archival materials and information on archives throughout Germany (collected from the German Digital Library). Using the concepts of MemO, the KG captures key metadata about resources, such as their source, creator, element type, or key themes/concepts. Vocabularies and concepts are harmonised (using standard data such as Wikidata or GND), which makes it possible to establish cross-connections between previously unrelated objects. For example, material from the Niedersächsisches Landesarchiv represented in Archivportal-D can be linked conceptually with other references to similar collections, e.g., in the EHRI portal. The ultimate goal is to create the NFDI4Memory KG as an index of the consortium's research resources and services in a shared space with a single point of entry. In the future, the KG will include more data, enabling semantic linking across sources and improving interoperability, new research exploration, and a FAIRer treatment of research data in the historical sciences and federation across the NFDI consortia.