Research outputs are no longer just journal articles. Datasets, software, models, protocols, and other non-textual outputs are increasingly recognised by funders, publishers, and ranking frameworks as first-class research contributions. A modern Research Information Management System (RIMS) treats them accordingly: each output is captured, attributed, and surfaced in the institutional record, not relegated to a separate spreadsheet or repository the rest of the institution cannot see.
Why datasets matter to a research office
Three converging trends have made dataset management an institutional concern, not just a librarian's:
- Funder mandates. Major funders — including European framework programmes, NIH, NSF, and Wellcome — require research data management plans and, increasingly, open availability of underlying data.
- Open science assessment. Research assessment frameworks (REF impact case studies, CoARA principles, Plan S) reward open, reusable outputs of all kinds.
- Discoverability and credit. Datasets with persistent identifiers (DOIs) and proper metadata are cited, counted, and contribute to the institution's research profile.
What a research dataset entry should contain
A defensible institutional record of a dataset goes well beyond a title and a download link. It needs:
- A persistent identifier (typically a DOI minted by a registry such as DataCite or Zenodo).
- Authorship and contributor roles, with ORCID iDs where available.
- A clear licence (Creative Commons or equivalent) covering reuse.
- Metadata describing the dataset's scope, format, and version.
- Linkage to the publications, projects, and grants the dataset supports.
- Where physically stored (institutional repository, domain repository, third-party platform).
The role of a RIMS
A RIMS is not a data repository — but it is the institutional system of record that points to where datasets live and how they relate to everything else. By ingesting metadata from DOI registries and federated repositories, a RIMS attributes each dataset to the right researchers and units, links it to the funded project it came out of, and surfaces it on researcher and unit profiles alongside the publications. The companion treatment of how a RIMS draws together global sources is covered in How a RIMS ingests data from five global sources.
Datasets and the single source of truth
When publication and dataset records live in separate systems, the institution cannot answer simple questions: "how many datasets did this researcher produce?", "which projects yielded both papers and reusable data?", "what is our open-data share?". When both sit in a reconciled RIMS, those questions are reports, not investigations. This is the principle we describe in Building a Single Source of Truth for Research Data.
Aligning with FAIR principles
The FAIR principles — Findable, Accessible, Interoperable, Reusable — are the de-facto standard for research data management. A RIMS supports each pillar: Findable through persistent identifiers and metadata, Accessible through clear licensing and links to the storage location, Interoperable through structured metadata that other systems can consume, and Reusable through provenance, version, and licence information attached to every record. This complements rather than replaces a specialised data repository.
How this differs from a repository
A repository stores the file; a RIMS records the institution's relationship with the work. The same dataset can be deposited in Zenodo, DataCite, or a discipline-specific archive, and still be tracked institutionally through the RIMS. The institution gets credit, the researcher gets attribution, the funder gets compliance evidence — all without forcing the dataset to live in a single system. See RIMS vs Institutional Repository for more.
Frequently asked questions
Does our RIMS need to host the datasets themselves? No. It tracks them — the dataset can live wherever is most appropriate technically and legally.
How are datasets cited? By DOI, which makes them counted and attributable through the same citation infrastructure used for publications.
What about software and code? The same model applies. Software outputs with DOIs (typically via Zenodo or similar) are recorded as research outputs alongside datasets and publications.
Where to start
If datasets and other non-textual outputs are not yet first-class records in your institutional system, that is the first gap to close. Discover RIMS captures publications, datasets, software, and other research outputs in one reconciled record, with links to the people, projects, and funders behind each — so the institution can credibly describe everything it produces, not just the papers.