Writing a Data Management Plan
The Data Management Plan (DMP) is a concise document that helps organize and anticipate all stages of the data life cycle. For each dataset, it explains how the data will be managed over the course of a project or within a research structure, from creation or collection through to sharing and archiving.
The DMP is a living document, so it is important to version it. For a research project, depending on the funder, up to 3 versions may be required: 6 months after the project starts, at mid-project, and at the end of the project. For research structures, it is important to keep the DMP up to date so that it reflects the structure's practices as accurately as possible.
The different types of DMP
Three types of data management plans can be distinguished:
- Project DMP: documents good data management practices over the course of a research project. Most often it is a project deliverable (see funder requirements).
- Entity DMP: documents good data management practices of a research structure (units, platforms, or research infrastructures). It offers a more global description of research practices than the Project DMP and is designed for the long term. Elements of the Entity DMP can be used to complete a Project DMP. Entity DMPs may be required by institutions, storage infrastructures, or funders.
- SMP (Software Management Plan): a management plan that documents FAIR practices in the development of research software.
Writing a DMP
What skills? What team?
Writing a DMP is a collective effort. It requires a coordinator within the project team, and sometimes within each work package. Involvement of the research teams is essential, since it will be necessary to describe precisely how data are collected, processed, and analyzed. You will also need support from an IT department to address aspects related to data storage, transfer, and security.
Depending on the project, legal experts may be consulted for questions relating to data protection (valorization, personal data, etc.). Finally, documentation and archives professionals can help you with the description, documentation, sharing, and archiving of your data. The LabRADoR team brings together all the skills needed to support you in writing a DMP; do not hesitate to contact us.
What tools?
To write a DMP, a simple word processor may be enough. Note, however, that dedicated software exists, which most often offers collaborative writing features. It also gives access to different templates or lets you create your own. Some of these tools are connected to repositories, databases, or APIs, which makes it possible to import information from external sources or export information, avoiding duplicate data entry; in that case we speak of Structured DMPs.
In France, the DMP OPIDoR tool is maintained by INIST-CNRS. It supports the templates of the main French and European research funders and offers connections with many French and international repositories. Note that the LabRADoR team can support you with this tool:
- We have an administrator account, which makes it easier to support the writing and review of your DMP
- We have written "Sorbonne University" recommendations that you can display to guide you as you write
- We offer training sessions or individual support to help you get started with DMP OPIDoR
Note, however, that other solutions exist for writing your DMP: DMPOnline, DSWizard, etc. In all cases, the LabRADoR team is available to help you get started with the software and with writing the DMP.
Classic DMP vs. Structured DMP?
In a "Classic" DMP, you describe your research practices mainly through free text, whether in a word processor or in software dedicated to writing DMPs. The classic DMP offers great freedom in writing, but everything must be completed manually, and at the end you will have an unstructured text document that is difficult to reuse or compare.
The Structured DMP (or maDMP, for machine-actionable DMP) is written using dedicated software. You describe your research practices through both free text and controlled lists. It offers auto-complete features through the use of persistent identifiers.
In most cases, it has an API that enables the transfer of information (for example, from an Entity DMP to a Project DMP). Its structured format facilitates its reuse and transfer from one software application to another.
The structured DMP template “Science Europe” available on DMP Opidor is becoming increasingly widely adopted in France. Its use is notably recommended by the ANR and the network of French funding agencies. This template supports:
- project data management plans and software data management plans (via the “software” data type)
- entity data management plans
DMP templates
A DMP template is a framework consisting of questions grouped by theme. It is designed to help you articulate your best practices in research data management. Templates are most often organized according to the data lifecycle or the FAIR principles. They may be created by funding agencies or institutions. Funding agencies recommend using their templates, but this is not mandatory if a specific template is better suited to the project’s data.
The structured DMP template “Science Europe” available on DMP Opidor is becoming increasingly widespread in France. Its use is notably recommended by the ANR and the network of French funding agencies. This template supports:
- project data management plans and software data management plans (via the “software” data type)
- Entity data management plans
However, other models exist, notably for the European programs Horizon Europe and ERC