Mapping the NARCIS aggregated domain to the OpenAIRE CERIF model

Abstract

The main entities in the NARCIS Portal (www.narcis.nl) domain consist of the following five main "collections": - Organizations, Researchers, Research projects, Publications and Datasets. The Publications and Datasets are being aggregated from different institutional PMH repositories. While the others originate from a relational database, called Nederlandse Onderzoeks Databank (NOD) that are actively maintained by DANS itself. All the Publications metadata comply to a national standard preserved by ‘EduStandaard'. This makes it easier to understand and interpret the collected metadata for an aggregator. Theoretically, all harvested 'collections' are represented as core CERIF entities and should easily be mapped to OpenAIRE CERIF entities. However, to be able to meet with the OpenAIRE Advance Guidelines for CRIS managers we had to re-classify, map and transform the existing model.

In this lightning talk I will focus on some of the issues we ran into while converting these entities:

● Mapping ‘EduStandaard' and NARCIS domain specific vocabularies to OpenAIRE (COAR). Different hierarchy of datatypes known only as 'subtypes' in the NARCIS model/domain;

● Name Identifiers: However deprecated (no resolver), the Dutch DAI is still in use to identify researchers and link researchers to their works and organizations;

● NARCIS as an aggregator: How do we deal with (local) identifier-types aggregated from different sources? We need to map federated Identifiers (Persons, Journals, etc.) to local ones to meet the PMH 'referential integrity' mandated by OpenAIRE guidelines;

● Linking research output to research projects.