How do I decide what data to keep?

Deciding which data to keep will differ from project to project but considering these questions can help when making decisions on what data to keep:

  • Does the data underpin published research findings?
  • Is there a funder or Imperial policy requirement to retain the data?
  • Are there any legal or contractual reasons to retain the data?
  • Does the data have potential for re-use in future research?
  • Is the data unique or difficult to replicate?
  • Does the data have scientific or historical value?

If your answer to one or more of these questions is yes, you will most likely want to keep the data beyond the end of your project.

Do I need to keep everything?

It may not be practicable or desirable to retain all data collected or generated beyond the end of a project. Retaining everything can make it harder to locate relevant files and for potential users to identify those most suitable for their needs. Preserving data can also involve significant costs, including not only storage but also the time and resources needed to manage, curate, and maintain it over the long term. The environmental impact should also be considered, as long-term preservation and access contribute to energy consumption, carbon emissions, water use, and electronic waste.

Consider whether all versions of a dataset need to be preserved. Files with no continuing evidential, legal, administrative, or analytical value, and which are not required for audit, regulatory, or retention purposes, may be disposed of. Examples include temporary or ephemeral files, unnecessary duplicate copies, and superseded working versions.

In some cases, particularly where large volumes of data are generated, preserving research data may not be cost-effective. For example, where data are generated through simulations or modelling, it may be preferable to preserve the code, software environment, and documentation needed to recreate the data rather than retaining the data themselves.

However, research data should not be disposed of until retention requirements, potential long-term value, and opportunities for future reuse have been carefully assessed.

How long should I keep my data?

Some data may be required by Imperial policy, funder policy, or even by law to be kept for a set period of time before it can be deleted. Imperial's Research Data Management Policy states that shareable research data should be retained for a minimum of 10 years, unless a funder requires a longer retention period. Some funders have data management and sharing policies which specify retention periods for datasets generated from funded projects.  See our web page What does my funder require? for brief summaries and links to individual funder data policies.

What should I do with my data at the end of my project?

The most effective and sustainable way to ensure the long-term preservation of research data is to deposit them in a trusted data repository. Imperial provides its own research data repository, Helix, which supports the preservation, discovery, and sharing of research data. In addition, there are many specialist repositories designed to meet the needs of research communities and data types. Guidance on selecting an appropriate repository is available on our Finding a research data repository webpage.

How should I dispose of research data?

Files that are not required for long-term preservation should be disposed of securely at the end of the project. Extra care should be taken when disposing of files that contain sensitive or confidential information, as simple deletion alone is not sufficient. For help with disposing of confidential data contact the ICT Service Desk.

Non-digital data

For help with archiving or disposing of non-digital research data contact the College Archives and Corporate Records Unit.

Additional resources:

The following resources were consulted drafting this guidance: