Authors: Johanna Laiho-Kauranne & Iida Lehikoinen, CSC

Introduction

Machine-actionable data management plans (maDMPs) represent a shift from static documentation to living, reusable research infrastructure. By structuring information for machine processing, maDMPs enable the use of AI to extract, connect, and reuse up-to-date data management information across systems. This description shows how structured, semantically enriched and integrated maDMPs deliver real value to researchers, far beyond meeting funder or organisational requirements defined in the Finnish national reference data model for maDMPs. The model has been developed within the OSTrails EOSC-INFRA project through wide national collaboration.

Built on the RDA maDMP standard, OSTrails Commons, and national requirements, the Finnish model provides a structured way to describe datasets, workflows, and software throughout the research lifecycle. Utilising persistent identifiers (PIDs) and ontologies, it enables DMP content to be reused, connected and automatically exchangeable between systems.

The data model is generic for all scientific disciplines, and it has been reviewed against the needs of SSH and ENVI research projects (status 07/2026).

Implementation 

The maDMP reference data model is tool agnostic, but it requires the tool to support certain machine-actionable structures. The data model has a total of 249 fields, but only 14* mandatory fields in its minimal implementation. The data model can be tailored to the needs of each organisation and scientific discipline by including and excluding sections and fields as appropriate. The model is intentionally broad to support this flexibility.

As this is a reference data model, it does not need to be followed exactly, and organisations can leave out mandatory fields, add new fields, or modify data types in the reference data model. 

* If the project uses data, the dataset section is also mandatory and includes 13 mandatory fields for each dataset described.

How to start

Organisation's data support/DMP tool admin

  1. Map current data model with national reference data model
    1. Review current maDMP / DMP data model and find corresponding fields for each
    2. Check national DMP requirements and organization's or funder's criteria
  2.  Start utilising the field names from reference data model in your DMP tool 
    1. Also utilise data types and cardinalities related to each field
    2. Explore which PIDs could be immediately used (ORCID, ROR, DOI, URN, ...) and which are needed to build interoperability between systems
    3. Collect feedback from researchers
  3. Create guidance
    1. Guidance for researchers in using the new maDMP template to support with changes
  4. Start expanding and improving the data model
    1. Expand organisation's data model with other relevant fields (especially mandatory ones) from the reference data model
    2. Use previous feedback to improve your data model
    3. Update guidance and maintain backlog for development needs
    4. Collect further feedback from researchers including also PIs and supervisors
  5. Repeat step 4 further as needed
    1. Further points to consider in developing your organisation's data model
      1. Do some scientific fields have specific needs that are not supported yet?
      2. Does your data model support consortium needs?
      3. Refine data model by type of research project (external/internal funded research project, student, other)

If you are interested, you can join a co-development pilot with CSC to test tools that support maDMP structures (ARGOS, DSW, FAIRWizard, DMPTuuli, DMPTool, RDMO) before starting to utilise the data model in your organization.

Researcher

  1. Read your organisation's guidance for maDMPs and DMP tool and start familiarising with your organisation's data model
  2. Use the new data model in your next project's DMP
    1. Make sure that you use PIDs as much as possible to enable extracting information automatically and to reduce your manual work in writing DMP
    2. Utilize metadata information directly from data repositories when re-using data
  3. Give feedback about your experiences with maDMPs to further develop the data model or DMP tool
  4. If you have external requirements for a maDMP data model, contact your organisation's data support and communicate your needs

Benefits of maDMP

How the researcher and research project benefit from maDMPs

For researchers, starting to utilise maDMPs means that information entered once can flow between systems, and stay consistent throughout the project. Instead of writing lengthy descriptive plans, researchers can create dynamic DMPs that link data, methods, services and outputs to repositories. maDMPs enable pre-filling information for resource applications, which can reduce both manual work and time required for the process. Overall, maDMPs improve data quality, support metadata reuse, support resource planning, and allow seamless information exchange between services while promoting FAIR data principles.

Moving from narrative DMPs to structured maDMPs helps researchers spend less time on administration and more on their research, while enabling smarter discovery of optimal and accessible services. Realising these benefits, however, requires a shift in practice. Familiarising oneself with and actively using PIDs and shared semantic structures is a key step towards creating living and reusable research infrastructure.

How the organisation can benefit from the use of maDMP reference data model

By using maDMPs, organisations gain access to structured information about the research conducted by their researchers, including project information, funding sources, responsibilities within projects, and researcher support needs. This enables better visibility into ongoing work and clearer insight into organisational needs. In addition, maDMPs can support organisations in managing intellectual property rights (IPR) by improving visibility into data ownership and usage conditions, as well as offering a basis for guidance on IPR-related matters. They can also contribute to handling contracts and liabilities by clarifying roles, responsibilities, and obligations linked to research outputs and data.

With this information, organisations can enhance multi-annual resource planning, allocate support services more effectively, and help researchers identify relevant research opportunities. Overall, maDMPs provide a clearer, organisation-wide understanding of research activities and researcher needs. However, the level of benefit depends on how interoperable the organisation’s internal systems are. It is also influenced by which PIDs are available and used consistently across systems and how well interfaces are designed to connect researcher-provided information in the DMP with organisational systems (e.g. codes identifying projects, decisions and declarations).

How the research services benefit from the use of maDMPs 

Research services can benefit from the improved clarity provided by the structured maDMP data model in the review process, since all DMPs would have a harmonised format. In addition, maDMPs provide earlier information about resource needs of different research projects. maDMPs can deliver information about the volume of resources, the expected period during which resources are needed, when the need occurs, the purpose for the resources, and the required protection or security level of the resources needed. With this information, it is possible to suggest better alternative resources for projects as well as plan for future use of resources. This applies to organisations' internal research services as well as external services (e.g. CSC and services from EOSC nodes).  

How the funders can utilise the maDMPs

maDMPs and the reference data model allow funders to receive clear overview of DMP content instead of solely relying manual review. Assessing whether DMPs follow funder requirements (e.g. Open Science and FAIR principles) will become easier with harmonised formats, and the structured data is easier to use in analytics. Many maDMP tools also have a FAIRness checking system, which could be used to communicate compliance information to funders in a simple and effortless way. 

maDMPs can also support funders in assessing the impact of their funding, as maDMPs used during the reporting stage may include DOIs for publications, datasets and software resulting from the funded research.

Further development

We welcome feedback to support the further development of the maDMP reference data model. We are also eager to support organisations in adopting and utilising maDMPs and to learn from the experiences of both organisations and researchers. Insights gained through implementation can help us share best practices and further improve the reference data model and use of maDMPs in services and across the broader research ecosystem.

  • No labels

Korkeakoulutuksen ja tutkimuksen yhteistyöalustan tietosuojaseloste