4. Research Context¶
4.1 Uses, Users and Expected Outcomes¶
This section sets out the range of uses the Data Director is designed to support, the users it serves, and the expected outcomes of each use. The primary outcome is to deliver the vision set out in Section 3.2: enabling any researcher to prepare and deposit research data for publication in alignment with the FAIR principles, regardless of their expertise or institutional support.
A use describes a specific situation in which a real user interacts with the tool to solve a real problem. Each use answers: who is doing what, why, and what happens as a result. The Working Group identified a range of uses for how the Data Director could support researchers and data professionals. Each has been assigned one of two statuses (Table 1):
Confirmed (green): in scope for this iteration of the Data Director. Confirmed uses are directly supported by a functional requirement in Section 7.1.
Future Consideration (yellow): addressing important gaps and challenges identified by the community, but beyond the scope of the current programme. These uses represent a ‘community wish list’ for future iterations of the Data Director. They are documented here to inform ongoing development; to signal the direction the community would like to see the tool evolve, and to provide a foundation for future Blueprint versions or related agentic AI tools in development.
The volume of future considerations reflects the richness of community input rather than a lack of ambition. Given the eight-week timeframe of the Blueprint programme, not all identified uses could be fully discussed, specified, or progressed to a point where they could be confirmed for this iteration. All future considerations are nonetheless valued contributions and are strongly encouraged for follow-up in future iterations of the Data Director or through related community-led initiatives.
Table 1. Uses, users and expected outcomes of the Data Director, showing confirmed uses for this iteration and community-identified future considerations.
| Use A specific interaction between user and tool |
User(s) Who initiates or benefits from this use |
Expected Outcome What the tool produces or enables |
Aligned Gap(s) The research challenge addressed in Section 3.1 |
Status In scope now, or identified for future development |
|---|---|---|---|---|
| 1. REPOSITORY SELECTION | ||||
| Search and recommend appropriate data repositories based on funder, institutional, journal, and disciplinary requirements | Researcher, Data Support Professional | A suitable repository is identified; conflicts between funder, institutional, journal and disciplinary requirements are flagged for human review; metadata recommendations for the selected repository are provided | Policy, Standards, Support | Confirmed |
| Set access conditions and assign a licence for data reuse, including open, restricted, and embargoed options | Researcher, Data Support Professional | Access conditions and licence are defined and attached to the dataset; open, restricted, and embargoed options are supported; funder or journal mandates are flagged where relevant | Policy, Governance | Future Consideration |
| Generate a private review link for manuscript peer reviewers to access an embargoed dataset before the associated journal article is published | Researcher, Data Support Professional | Peer reviewers can access the dataset under embargo without it being made publicly available; access is time-limited and auditable | Policy, Governance | Future Consideration |
| Monitor embargo status and alert the researcher when a dataset should be released to public access following journal article publication | Researcher, Data Support Professional | The researcher is notified when the embargo period has elapsed and the associated journal article has been published; release to public access is confirmed by the researcher before being enacted | Policy, Governance | Future Consideration |
| Check repository certification and trustworthiness, for example against CoreTrustSeal1 or FAIRsharing2 criteria | Researcher, Data Support Professional | Repository trustworthiness is surfaced using recognised schemes such as CoreTrustSeal and FAIRsharing; free certification options are prioritised over paid alternatives | Policy, Support | Future Consideration |
| Provide guidance on managing and sharing very large data files, including file compression, transfer methods, and upload requirements for discipline-specific repositories | Researcher, Data Support Professional |
Researchers are guided towards appropriate solutions for large file storage, transfer, and format conversion; barriers to sharing large datasets are reduced | Capacity, Support | Future Consideration |
| 2. METADATA AND DOCUMENTATION | ||||
| Create FAIR-aligned and discipline-appropriate metadata, including controlled vocabularies, ontologies, and community standards such as CARE principles where applicable | Researcher, Data Support Professional, Institution | A draft metadata record is generated using appropriate controlled vocabularies (standardised term lists) and ontologies (structured concept frameworks); the researcher reviews and approves before submission | Capacity, Standards, Support | Confirmed |
| Suggest appropriate vocabularies and ontologies | Researcher, Data Support Professional | Improved consistency, interoperability, and discoverability of data, enabling machine-readable metadata and supporting reuse across systems and communities | Standards, Capacity | Confirmed |
| Recommend appropriate open and non-proprietary file formats for a dataset to support long-term accessibility and interoperability | Researcher, Data Support Professional | Researchers are guided towards open, self-describing formats suitable for sharing and reuse; format recommendations are provided across the data publication workflow | Standards, Capacity | Confirmed |
| Validate and verify metadata quality against community best practices, recommendations and standards, producing records that support data reuse | Researcher, Institution | Harmonised, machine-interpretable metadata records that meet applicable community best practices and recommendations and support data reuse | Capacity, Standards, Quality and Provenance | Confirmed |
| Draft and review data documentation, including READMEs and data dictionaries, across the data publication workflow | Researcher, Data Support Professional | Clear documentation that enables others to understand and reuse the data | Capacity, Support | Confirmed |
| Map metadata across multiple schemas so the same dataset can be described in more than one standard, enabling deposit across different repositories without manual remapping | Researcher, Data Support Professional | Metadata is expressed in multiple schema formats simultaneously, reducing barriers to deposit across different repositories and improving interoperability without requiring the researcher to manually remap their record | Standards, Capacity | Confirmed |
| Enable consistent, reusable metadata across institutions and between different deployments of the Data Director, so data created in one system can be understood and used in another | Researcher, Data Support Professional | Existing metadata formats and schemas are reused; less data wrangling; easier interpretation of data across systems | Standards, Capacity | Confirmed |
| Detect and flag information in datasets that could identify individuals, either directly or indirectly, before publication | Researcher, Data Support Professional | Potential identifiers are flagged before the publication workflow proceeds; the researcher is alerted and guided on appropriate next steps; the tool does not attempt anonymisation or deidentification autonomously | Governance, Policy | Future Consideration |
| Check that variables use correct and consistent terminology to support future meta-analysis and data reuse | Researcher, Data Support Professional | Variable terminology is flagged where it may affect interoperability or reuse; researchers are guided towards community-agreed definitions | Standards, Capacity | Future Consideration |
| Generate or assist with Model Cards or Dataset Cards for machine learning research datasets | Researcher, Data Support Professional | A structured Model Card or Dataset Card is produced for machine learning datasets, improving transparency and reusability. Note: this use applies specifically to machine learning datasets | Capacity, Standards | Future Consideration |
| Create crosswalks between metadata schemas to support interoperability | Researcher | Datasets from different sources can be combined and compared, for example medical imaging data with geospatial data | Standards, Capacity | Future Consideration |
| Harmonise metadata schemas across repositories to improve consistency and FAIR alignment across institutional and regional systems | Data Support Professional | Interoperable repositories at institutional, regional and national level | Standards, Capacity | Future Consideration |
| Connect multiple Data Director deployments so datasets across institutional boundaries can be discovered and shared through a joined-up system, including data from third-party sources | Researcher, Data Support Professional | Data held across institutional boundaries and in external sources becomes discoverable and shareable through connected Data Director instances | Standards, Capacity | Future Consideration |
| 3. PERSISTENT IDENTIFICATION | ||||
| Support the assignment of persistent identifiers (PIDs) to datasets at the point of depositing data for publication; users are informed if the selected repository assigns PIDs automatically | Researcher, Institution | Datasets are citable and discoverable via persistent identifiers (stable, long-lasting references that remain valid if a dataset moves location); impact metrics can be collected; PID assignment remains the responsibility of the repository | Support, Standards | Confirmed |
| Support PID assignment at variable or file level, not only at dataset level, to enable more granular citation and traceability | Researcher, Data Support Professional | Individual variables or files within a dataset are persistently identifiable and citable; granular provenance and credit attribution is enabled | Standards, Quality and Provenance | Future Consideration |
| Attach machine-readable licence and access policies to datasets so that permitted uses travel with the data and can be understood by both humans and automated systems | Researcher, Data Support Professional | Licence and access conditions are machine-readable and travel with the dataset as a persistent, FAIR-aligned policy object | Governance, Standards | Future Consideration |
| 4. DATA MANAGEMENT PLANNING | ||||
| Check that research data publication matches what was committed to in the Data Management Plan (DMP), where one exists | Researcher | Research data is published in line with the original plan; any discrepancies are flagged for human review; the tool functions without a DMP where one is not available | Capacity, Policy | Confirmed |
| Draft, review and update DMPs, incorporating FAIR and CARE principle considerations; CARE principles relate to the governance of Indigenous data. | Researcher, Data Support Professional | A DMP that identifies FAIR and CARE gaps, meets funder and institutional requirements, and is kept up to date throughout the project | Capacity, Policy, Governance | Future Consideration |
| Update the DMP to reflect research data as published, where funders require this | Researcher, Data Support Professional | The DMP accurately reflects the final data publication; funder compliance is maintained post-publication | Capacity, Policy | Future Consideration |
| 5. METADATA REMEDIATION AND IMPROVEMENT | ||||
| Recommend improved metadata for already-published datasets and supplementary materials, including preliminary and unpublished datasets, for researcher review and action | Researcher, Data Support Professional | Existing datasets become more discoverable, citable, and reusable; metadata is improved against FAIR principles and community best practices | Capacity, Standards | Confirmed |
| Use automated data processing to extract, reformat and improve existing metadata records, applying the same standards-selection process used for new data | Researcher, Data Support Professional | Existing metadata is cleaned and updated using the same standards-selection process as new data | Capacity, Standards | Future Consideration |
| 6. FAIR ASSESSMENT | ||||
| Connect to recognised FAIR evaluation tools to assess how well produced metadata meets community standards, returning a score or maturity level to guide improvement | Researcher, Data Support Professional, Institution | A FAIR evaluation score or maturity level is returned by the connected evaluation tool; researchers can understand, track, and improve their metadata against recognised criteria | Standards, Quality and Provenance | Future Consideration |
| Evaluate how FAIR-aligned already-published datasets are using recognised tools such as FAIRassist,3 and provide improvement recommendations | Researcher, Data Support Professional | The FAIRness of existing published outputs is assessed against community benchmarks; improvement recommendations are provided | Standards, Quality and Provenance | Future Consideration |
| 7. PROVENANCE AND TRACKING | ||||
| Track the provenance chain of AI agent actions on data and metadata | Researcher, Data Support Professional, Institution | All agent actions are recorded and can be verified and validated | Quality and Provenance, Standards | Confirmed |
| Produce provenance records for existing datasets and enable them to be cited and connected to related outputs, so data sources are traceable and credit is attributed | Researcher, Data Support Professional | Data sources are traceable; credit is attributed to the original dataset | Quality and Provenance, Standards | Future Consideration |
| Assign persistent identifiers and provenance records to individual data variables so each carries full context including definition, units, and origin, supporting human and machine use | Researcher, Data Support Professional | Individual variables are identifiable, citable, and carry full contextual provenance including definition, units, properties, and relationships | Quality and Provenance, Standards | Future Consideration |
| Digitally sign large language model prompts and verified question and answer pairs to record ownership and provenance of AI-generated content | Researcher, Data Support Professional | AI-generated content is traceable to its source prompt; ownership is recorded with persistent identifiers; verified human-approved outputs are signed and auditable | Quality and Provenance | Future Consideration |
| 8. DISCOVERY AND LINKING | ||||
| Link datasets to related research outputs such as journal article publications, people, grants and software at the point of publication | Researcher, Data Support Professional, Institution | Persistent links are established; richer reporting across research assets; more contextual discovery of outputs | Standards, Support | Confirmed |
| Discover related datasets relevant to a research project | Researcher | Relevant data is surfaced quickly without relying on general-purpose search engines; discovery does not affect existing metadata | Standards, Support | Confirmed |
| Include data citations in related journal article publications and record the relationship between a dataset and its associated publication in the metadata | Researcher, Data Support Professional | Datasets are linked to related journal article publications using metadata properties such as IsReferencedBy and IsSupplementTo; where the article cannot be updated, the relationship is recorded in the dataset metadata | Standards, Support | Future Consideration |
| Record deposit in institutional systems such as Current Research Information Systems (CRIS) | Researcher, Data Support Professional, Institution | Deposit is recorded in the relevant institutional Current Research Information System (CRIS), supporting institutional reporting on research outputs and activities | Policy, Support | Future Consideration |
| Monitor access requests, downloads and citations for published datasets and provide researchers with metrics on data reuse | Researcher, Institution | Researchers receive regular updates on how their data is being accessed and reused; metrics support recognition of data outputs within the current reward system | Standards, Support | Future Consideration |
| Link datasets to protocol repositories to enhance data replicability and methodological transparency | Researcher, Data Support Professional | Datasets are connected to the protocols used to generate them; replicability is supported through persistent links to methodology | Quality and Provenance, Standards | Future Consideration |
| Find journal article publications and linked datasets on a given research topic | Researcher | A bibliography of relevant papers and associated datasets is produced | Standards, Support | Future Consideration |
| Build knowledge graphs, for example from nanopublications | Researcher | Knowledge is shared and managed in a structured, machine-readable form | Standards | Future Consideration |
| 9. PRIVACY AND COMPLIANCE | ||||
| Carry out privacy due diligence for research involving human subjects at the point of publication | Researcher, Data Support Professional | Consent documents and data use agreements are checked; identifiers are removed; ethics approval documentation is linked; sharing is confirmed as legally compliant | Governance, Policy | Future Consideration |
| Manage data access automatically in line with applicable data protection legislation, for example restricting access to sensitive datasets, subject to explicit human approval | Data Support Professional | Data access is managed in line with legal requirements; no access restriction is enacted without human confirmation | Governance, Policy | Future Consideration |
| Monitor compliance and recommend remediation where data does not meet national or institutional policies | Data Support Professional | Non-compliance is identified; fixes are suggested, such as changes to metadata, file formats, or updates to related repository records. | Governance, Policy | Future Consideration |
| 10. INDIGENOUS DATA | ||||
| Ensure data reflects Indigenous context and aligns with CARE principles at the point of publication | Researcher, Data Support Professional | Data is handled in a way that is not biased; Indigenous data rights and protocols are respected | Governance, Standards | Future Consideration |
| 11. AI AGENT GOVERNANCE | ||||
| Establish digital identities for AI agents and manage their rights and permissions | Researcher, Data Support Professional | Humans can create agents with clearly defined, responsible rights; agents can return access and permissions to humans; data transfer is policy-driven and restricted appropriately | Governance, Quality and Provenance | Future Consideration |
| 12. COMMUNITY OVERVIEW | ||||
| Provide an overview of data practices within a research community | Researcher, Data Support Professional | A dashboard showing the metadata standards and most-used repositories within a community | Standards, Support | Future Consideration |