Abstract:
Cultural heritage data, encompassing tangible, intangible, and natural assets as defined by UNESCO, has seen growing digitization efforts globally. While several initiatives and platforms facilitate access to Cultural Heritage content, the Linked Open Data (LOD) ecosystem still lacks a dedicated and curated index for Cultural Heritage datasets and resources. This absence limits the discoverability, accessibility, and reuse of valuable Cultural Heritage data.
This study introduces \texttt{CHeCLOUD}—the Cultural Heritage Linked Open Data Cloud—a topical sub-cloud within the broader LOD Cloud, designed to improve the FAIRness (Findability, Accessibility, Interoperability, Reusability) of Cultural Heritage datasets. The goal is to provide a sustainable, centralized reference for Cultural Heritage LOD datasets and assess their quality through a FAIR-aligned lens. The selection of the datasets to be included in \texttt{CHeCLOUD} followed a three-phase methodology inspired by systematic literature review guidelines, adapted for dataset discovery. It includes: (1) structured identification of Cultural Heritage datasets from the LOD Cloud and external sources; (2) FAIRness evaluation using a mapping framework between data quality and linked data principles as well as the KGHeartBeat quality assessment tool; and (3) maintenance, via continuous updates and a feedback mechanism. The methodology ensures transparency, reproducibility, and domain-specific relevance.
Besides detailing the proposed resource, datasets are assessed in terms of FAIR-ness by aligning FAIR principles and linked data quality dimensions. As a result, Reusability emerges as the strongest dimension, primarily due to consistent licensing and provenance metadata. Accessibility also scores relatively high, while Findability and Interoperability reveal notable gaps, especially regarding metadata richness, URI dereferenceability, and vocabulary reuse.
CHeCLOUD fills a critical gap in the LOD ecosystem by offering, for the first time, a structured, FAIR-aligned index of Cultural Heritage datasets. The findings highlight both the current strengths and areas needing improvement in Cultural Heritage data publication practices. The proposed methodology and assessment framework can be generalized to other domains, supporting broader efforts to enhance data FAIRness across Linked Data ecosystems.
