Building on the previous post which described our IDMB-inspired basic metadata model, this post will present some more tangible results of our work so far.
Toward a standard metadata schema for research data
We thought the best place to start in developing our metadata schema would be to assess the strengths and weaknesses of a bunch of different metadata specifications. This exercise would help inform the development of our generic schema for describing research data.
- The top level relates to the Discoverability of the Resource and could be based around the 15 Dublin Core elements
- The second level is contextual and essentially covers the elements in cerif (notably; outcomes – publications, patents etc, funding; research council grant, person; project team members, organisation- Uni, collaborators.
- Finally a specific level giving the minutiae.
Levels 1 and 2 are fairly well covered by default EPrints metadata. The rather more vague third point demands more consideration - which 'minutiae' should we provide? A mapping (or crosswalk) of candidate schema is attached below, and includes an early (draft) version of the proposed RDE metadata schema (mapped to EPrints default). This draws on elements from DataCite, INSPIRE, DDI and DataShare schema.
Designing an interface for research data eprints
The image below shows a mockup of a generic data eprint, and the way we hope to present the metadata above. This design illustrates how data, documentation and the various levels of metadata will appear to a user. Significant changes are required to the way in which the items (i.e. files) are displayed, in order to make the meaning of the various elements of a complex collection intelligible.
A trial implementation is in production and will likely be refined and improved as we progress.