Conference paper Open Access

Context-Driven Discoverability of Research Data

Miriam Baglioni; Paolo Manghi; Andrea Mannocci

Research data sharing has been proved to be key for accelerating scientific progress and fostering interdisciplinary research; hence, the ability to search, discover and reuse data items is nowadays vital in doing science. However, research data discovery is yet an open challenge. In many cases, descriptive metadata exhibit poor quality, and the ability to automatically enrich metadata with semantic information is limited by the data files format, which is typically not textual and hard to mine. More generally, however, researchers would like to find data used across different research experiments or even disciplines. Such needs are not met by traditional metadata description schemata, which are designed to freeze research data features at deposition time.

In this paper, we propose a methodology that enables “context-driven discovery” for research data thanks to their proven usage across research activities that might differ from the original one, potentially across diverse disciplines. The methodology exploits the collection of publication–dataset and dataset–dataset links provided by OpenAIRE Scholexplorer data citation index so to propagate articles metadata into related research datasets by leveraging semantic relatedness. Such “context propagation” process enables the construction of “context-enriched” metadata of datasets, which enables “context-driven” discoverability of research data. To this end, we provide a real-case evaluation of this technique applied to Scholexplorer. Due to the broad coverage of Scholexplorer, the evaluation documents the effectiveness of this technique at improving data discovery on a variety of research data repositories and databases.

Please cite this paper as: Baglioni, M., Manghi, P. and Mannocci, A., "Context-Driven Discoverability of Research Data", In proceedings of the 24th International Conference on Theory and Practice of Digital Libraries (TPDL), Lyon, France, 2020. doi: 10.1007/978-3-030-54956-5_15
Files (358.2 kB)
Name Size
Context_driven_Discoverability_of_Research_Data___Revised.pdf
md5:605d29fe83bc3aaca7ed75298834b136
358.2 kB Download
67
95
views
downloads
Views 67
Downloads 95
Data volume 34.0 MB
Unique views 63
Unique downloads 91

Share

Cite as