A primer on managing your research data

“Open data and content can be freely used, modified and shared by anyone for any purpose”

Source: http://opendefinition.org

Tim Berners-Lee’s proposal for five star open data http://5stardata.info

4 Key steps in very approximate order — many of the steps can be done simultaneously.

1. Choose your dataset(s) — Choose the dataset(s) you plan to make open. Keep in mind that you can (and may need to) return to this step if you encounter problems at a later stage.
2. Apply an open license (see more info)
   • Determine what intellectual property rights exist in the data.
   • Apply a suitable ‘open’ license that licenses all of these rights and supports the definition of openness.
   • NB: if you can’t do this go back to step 1 and try a different dataset.
3. Make the data available — in bulk and in a useful format. You may also wish to consider alternative ways of making it available such as via an API.
4. Make it discoverable — post on the web and perhaps organize a central catalog to list your open datasets.

Source: https://okfn.org/opendata/how-to-open-data/

Image courtesy of http://aukeherrema.nl CC-BY

(Not for the research funder, but for life we make data management plans)

• Make your research easier
• Stop yourself drowning in irrelevant stuff
• Save data for later
• Avoid accusations of fraud or bad science
• Write a data paper
• Share your data for re-use
• Get credit for it

• The principal investigator – ultimately responsible for the data and for data management
• Researchers, research assistants and/or data managers – involved in day-to-day data management
• The institution’s management – draft and enforce data policies; raise data awareness
• The institution’s research office consisting of library, IT and legal services – provide external data, tools, secure storage and access; expertise on rights management and ethics, data citation, metadata, access and licenses, funder requirements; raise data awareness
• Research funders – encourage good data practices; invest in data infrastructure; raise data awareness
• Project partners in academic and other research institutions as well as commercial partners
• Academic publishers – impose requirements on the availability of data underlying submitted and/or published papers; provide identifiers to cite papers and link to related data
• Research data repositories – long-term preservation of data long term; provide persistent identifiers and data discovery service

• The data needed to validate results in scientific publications (minimally!).
• The associated metadata: the dataset’s creator, title, year of publication, repository, identifier etc.
  • Follow a metadata standard in your line of work, or a generic standard, e.g. Dublin Core or DataCite, and be FAIR.
  • The repository will assign a persistent ID to the dataset: important for discovering and citing the data.
• Documentation: code books, lab journals, informed consent forms – domain-dependent, and important for understanding the data and combining them with other data sources.
• Software, hardware, tools, syntax queries, machine configurations – domain-dependent, and important for using the data. (Alternative: information about the software etc.)

Basically, everything that is needed to replicate a study should be available. Plus everything that is potentially useful for others.

Check out case studies and overviews of research data repositories, funder policies on data sharing and researchers attitudes towards data sharing in the Open Research Data section of the Open Science Monitor, commissioned by the European Commission Directorate-General for Research and Innovation

Image courtesy of http://aukeherrema.nl CC-BY

A Data Management Plan (DMP) is a brief plan to define:

• how the data will be created
• how it will be documented
• who will be able to access it
• where it will be stored
• who will back it up
• whether (and how) it will be shared and preserved.

DMPs are often submitted as part of grant applications, but are useful whenever researchers are creating data.

Common themes in DMPs

• Description of data to be collected / created (i.e. content, type, format, volume...)
• Standards / methodologies for data collection & management
• Ethics and Intellectual Property (highlight restrictions on data sharing e.g. embargoes, confidentiality)
• Plans for data sharing and access (i.e. how, when, to whom)
• Strategy for long-term preservation

More elaborate DMP

Scientific research data should be easily:

1. Discoverable - Are the data discoverable and identifiable by a standard mechanism e.g. DOIs?
2. Accessible - Are the data accessible and under what conditions e.g. licenses, embargoes?
3. Assessable and intelligible - Are the data and software assessable and intelligible to third parties for peer-review? E.g. can judgements be made about their reliability and the competence of those who created them?
4. Useable beyond the original purpose for which it was collected - Are the data properly curated and stored together with the minimum software and documentation to be useful by third parties in the long-term?
5. Interoperable to specific quality standards - Are the data and software interoperable, allowing data exchange? E.g. were common formats and standards for metadata used?

 

• The principles of good research conduct hold for all of us, across disciplinary boundaries.
• Data management is all in a day’s work.
• Planning and reflection are more important than the plan – but write the DMP and keep it up to date.
• Planning data management is teamwork.
• Think about the desired end result and plan for this.
• Decisions made early affect what you can do later.

• MANTRA Research Data Management Training - a free online course for those who manage digital data as part of their research project
• Briefing papers, How-to Guides and Checklists, Tools, Case studies, Standards, etc from DCC, Digital Curation Center
• Essentials 4 Data Support - an introductory course for those who (want to) support researchers in storing, managing, archiving and sharing their research data