[Paper Review] Linked open data per la valorizzazione di collezioni culturali: il dataset mythLOD

Valentina Pasqual, Francesca Tomasi|arXiv (Cornell University)|Jan 1, 2022

Digital Humanities and Scholarship3 citations

TL;DR

This paper presents mythLOD, a linked open data platform that reimagines the Mythologiae digital art collection through semantic web technologies, transforming tabular data into a knowledge graph by formalizing expert-verified links between artworks and classical literary sources. The workflow includes data modeling, entity disambiguation, RDF conversion, and validation via competency questions and interactive visualizations—resulting in a reusable, extensible framework for cultural heritage interoperability and enhanced user exploration through an online catalogue and Aeneid-focused storytelling interface.

ABSTRACT

La rappresentazione formale dei metadati culturali è sempre stata una sfida, considerando tanto l’eterogeneità degli oggetti culturali quanto l’esigenza di documentare l’atto interpretativo esercitato dagli esperti. Questo articolo presenta una panoramica della ri valorizzazione della collezione digitale Mythologiae in formato linked open data. La ricerca mira a esplorare i dati di una raccolta di opere d’arte (Mythologiae) promuovendo le potenzialità del web semantico, concentrandosi in particolare sulla rappresentazione formale dell’associazione degli oggetti culturali alle fonti letterarie, così come realizzata dagli esperti, documentando anche le loro interpretazioni.
 Il flusso di lavoro è consistito nella definizione del modello di dati, nella pulizia e disambiguazione degli stessi, nella conversione (da struttura tabulare a grafo) e nell’attività di testing (in particolare la revisione degli esperti di dominio del dataset tramite competency question e visualizzazioni dei dati).
 Il risultato è la piattaforma mythLOD, che presenta il dataset insieme alla documentazione dettagliata della ricerca. Inoltre, la piattaforma ospita i due spazi di visualizzazione dei dati (il catalogo online e un esperimento di data storytelling sul caso studio dell’Eneide) che arricchiscono la documentazione del progetto come unità di test user-friendly per il dataset e costituiscono un ulteriore strumento di documentazione del progetto e di esplorazione della collezione.

Motivation & Objective

• To address the challenge of formalizing heterogeneous cultural metadata and expert interpretations in digital heritage collections.
• To enhance the expressiveness and reusability of cultural data by migrating from relational to semantic web models.
• To develop a replicable, standardized workflow for converting cultural collections into linked open data with semantic richness.
• To support user experience through interactive visualizations—online catalogue and thematic storytelling—enabling non-conventional access to cultural data.
• To demonstrate the value of linked data in connecting cultural objects with their literary and interpretative contexts, fostering new forms of knowledge discovery.

Proposed method

• Defined a custom data model based on domain expert analysis to represent both artworks and their literary associations with semantic precision.
• Performed data cleaning and entity disambiguation using controlled vocabularies and external alignment to standard knowledge bases.
• Converted relational tabular data into RDF triples using semantic web standards, establishing a graph-based knowledge representation.
• Integrated external linked data sources (e.g., Perseus Digital Library) to enrich metadata and enable federated querying.
• Validated the dataset through competency questions and expert review using purpose-built visualizations for semantic and syntactic consistency.
• Developed two interactive data visualization layers: a faceted online catalogue and a narrative storytelling interface centered on the Aeneid.

Experimental results

Research questions

• RQ1 How can expert-verified associations between visual artworks and classical literary sources be formally represented in a machine-processable format?
• RQ2 What methodological workflow ensures data quality and semantic expressiveness during the migration from relational to linked open data models in cultural heritage?
• RQ3 To what extent can linked open data and interactive visualizations enhance user access and knowledge discovery in digital art collections?
• RQ4 How can the formalization of interpretative acts be preserved and made accessible in a semantic knowledge graph?
• RQ5 What role do reusable, documented workflows play in enabling the replication of linked data projects across similar cultural heritage initiatives?

Key findings

• The mythLOD platform successfully transformed the Mythologiae digital collection into a linked open data knowledge graph, enabling machine-readable, semantically enriched interlinking between artworks and literary sources.
• The use of competency questions and expert review of visualizations confirmed the syntactic and semantic correctness of the dataset, ensuring high data quality.
• The online catalogue, enhanced with faceted filtering over metadata, literary sources, and interpretative acts, proved effective for data exploration and validation.
• The storytelling interface, which visualizes temporal and spatial dimensions of the Aeneid, demonstrated how linked data enables narrative-driven access to cultural collections.
• The entire workflow, including documentation and platform architecture, is openly released and designed for replication in similar cultural heritage digitization projects.
• The integration of external linked data sources (e.g., Perseus) enabled contextual enrichment and cross-referencing, significantly enhancing the dataset’s utility and discoverability.