Assessing Metadata Quality Across Helmholtz Data Providers: A Practical Approach for Harmonization

This poster presents results from the TF Harmony kick-off workshop and an accompanying large-scale metadata gap analysis conducted within the Helmholtz Metadata Collaboration (HMC). The work explores how publication metadata is currently structured across Helmholtz infrastructures and identifies both technical patterns and underlying contextual factors that influence metadata quality. The quantitative analysis is based on approximately 3 million metadata records harvested via OAI-PMH and focuses on three priority fields central to interoperability: identifier (dc:identifier), publication date (dc:date), and resource type (dc:type). Across these fields, simple and reproducible assessment criteria were applied to evaluate presence, representation, and standardization. The results show that while field availability is generally high, important inconsistencies persist, particularly in the use of persistent identifiers, controlled vocabularies, and standardized date formats. To better understand the reasons behind these patterns, a community workshop with representatives from 17 Helmholtz centres complemented the analysis. Structured breakout discussions were designed to capture participant perspectives on metadata practices, focusing on three dimensions:(1) constraints shaping metadata work,(2) semantic challenges and ambiguities, and(3) motivations and concerns regarding HMC support. The combined qualitative and semantic analysis of workshop notes shows that metadata challenges are not primarily perceived as purely technical issues. Instead, they are strongly shaped by organizational structures, resource limitations, legacy systems, and unclear responsibilities. These constraint-related factors were the most frequently observed themes, as also reflected in the Sankey-style synthesis of discussion results. At the same time, participants highlighted recurring semantic issues such as inconsistent terminology, free-text usage, and difficulties applying standards in domain-specific contexts. Participants expressed clear interest in improved harmonization, but emphasized the need for practical guidance, flexible approaches, and center-specific support, rather than prescriptive, one-size-fits-all solutions. The findings suggest that improving metadata interoperability requires addressing both technical standardization and the socio-organizational context in which metadata is created and maintained. By linking large-scale metadata analysis with community input, this work provides a grounded, evidence-based perspective on metadata quality across Helmholtz. The results inform ongoing HMC activities, including targeted consulting with data providers, iterative feedback loops, and the development of harmonization strategies. Ultimately, the work contributes to improving metadata consistency, strengthening connectivity in the Helmholtz Knowledge Graph, and enhancing the functionality of the FAIR Data Dashboard.