Giving Researchers Credit for their Data
Institutions are increasingly developing IRs (institutional repositories) to hold the data outputs of their researchers, helping to reduce the individual burden of data archiving. However, only a subset of the data produced is associated with publications and thus reliably archived, and much important data is never published, shared or re-used. This represents a loss of scientific knowledge, may lead to the repetition of research and wastes public money.
To incentivise researchers to share unpublished data, we aim to develop a simple ‘one-click’ process where data, metadata and methods detail are transferred from an IR to a relevant publisher platform for publication as a data paper. This can be peer reviewed, indexed to increase visibility, and recognised by the community as a formal research output. Subsequently, details of the paper can be fed back to the IR to enrich the original record.
The project is split into three phases to match the operation of the Jisc Data Spring programme.
Phase 1: Feasibility Study: Analyse publisher workflows and repository functions and perform a simple gap analysis. Develop a simple straw-man workflow between an IR and a Publisher that bridges this gap - to enable researchers to easily publish datasets as peer reviewed data articles. Survey the community of publishers and repositories to establish demand and interest in such a workflow.
Phase 2: Definition: Formalise these processes as an API definition (or extensions to existing APIs, e.g. SWORD) that has the potential to be used by other IR platforms (Fedora, DSpace, ePrints etc.) and publisher/editorial systems. We will publish a draft API and then hold a workshop open to all stakeholders, and seek input from the JISC Research Information Management Group.
Phase 3: Reference implementation: Develop a simple demonstration implementation between a data repository (such as the Oxford University Research Data Archive) and a data paper publisher (F1000Research). This would demonstrate the benefits, provide a reference example for other instances and, through the use of test-driven development, provide a reusable test suite. It is anticipated that the implementation would take the form of a bridging application that would have the potential to de deployed “in-the-cloud” supporting many-to-many transactions.