Looking at a digital research data archive - Visual interfaces to EASY

In this paper we explore visually the structure of the collection of a digital research data archive in terms of metadata for deposited datasets. We look into the distribution of datasets over different scientific fields; the role of main depositors (persons and institutions) in different fields, and main access choices for the deposited datasets. We argue that visual analytics of metadata of collections can be used in multiple ways: to inform the archive about structure and growth of its collection; to foster collections strategies; and to check metadata consistency. We combine visual analytics and visual enhanced browsing introducing a set of web-based, interactive visual interfaces to the archive’s collection. We discuss how text based search combined with visual enhanced browsing enhances data access, navigation, and reuse.

💡 Research Summary

The paper presents a comprehensive visual‑analytics study of the EASY digital research data archive, focusing on the metadata that describe deposited datasets. The authors first extract and clean key metadata fields—scientific discipline, depositor (person or institution), and access rights (open, restricted, or closed)—and then employ a series of visualizations to reveal structural patterns within the collection.

Using heat‑maps and treemaps, they illustrate the distribution of datasets across disciplines and show that fields such as life sciences, physics, and social sciences dominate the archive, with life sciences experiencing a particularly steep growth in the last five years. A network‑graph analysis of depositors highlights a small number of large universities and research institutes that act as hubs in specific domains, while individual researchers contribute mainly to smaller projects. This depositor network can inform collection‑development strategies by identifying key partners for outreach and support.

Access‑rights visualizations (pie charts and stacked bar charts) expose disciplinary differences in openness: physics datasets are overwhelmingly open (≈70 %), whereas social‑science datasets are more often restricted (≈55 %). These insights are valuable for policy makers seeking to balance open‑science objectives with privacy or commercial constraints.

The core contribution is a web‑based interactive interface that integrates text‑based search with the visual analytics described above. Built on D3.js and Leaflet, the dashboard allows users to type keywords, then instantly see matching datasets highlighted on maps, timelines, and discipline charts. Interactive filters (time sliders, discipline selectors, access‑rights toggles) enable dynamic narrowing of results, while zoom‑and‑pan, tooltips, and pop‑up metadata panels provide immediate access to detailed information without leaving the visual context.

To ensure metadata quality, the system incorporates an automated rule engine that flags missing mandatory fields, inconsistent date formats, and duplicate records in real time, allowing curators to correct errors on the fly. This dual focus on visual exploration and data‑quality monitoring demonstrates how metadata‑driven visual analytics can serve both end‑users and archive administrators.

The authors argue that such visual approaches serve three main purposes: (1) informing the archive about the current structure and growth trajectory of its collection, thereby supporting evidence‑based budgeting and infrastructure planning; (2) fostering collection‑development strategies by revealing dominant depositors and disciplinary gaps; and (3) checking metadata consistency to improve long‑term preservation and discoverability.

In the discussion, they note that visual analytics can uncover hidden trends, guide targeted outreach, and enhance the user experience compared with traditional list‑based search interfaces. They also outline future work, including machine‑learning‑based automatic classification of new deposits, user‑behavior analytics for personalized recommendations, and cross‑archive metadata linking to create broader, federated visualizations.

Overall, the study demonstrates that combining visual analytics with interactive browsing not only clarifies the internal composition of a research data archive but also actively improves data access, navigation, and reuse for a diverse stakeholder community.