Skip to main content

Describe data

Ensure that everyone can find and understand your research data

Accurately describing your data is essential in ensuring that you, and others who may need to use the data, can make sense of your data and understand the processes that have been followed during data collection, processing and analysis. Well-described data is more easily discoverable and verifiable.

Describing your data: metadata

Metadata is descriptive information about your data. When describing your data you’ll need to create a lot of the metadata yourself, such as ‘title’, ‘description’ and ‘creator’, while your computer system will create additional technical metadata about your files. All of this metadata is useful for understanding, managing, and preserving your data effectively and makes your data more findable and usable to other researchers if the data are published or shared. Find more information about metadata on the ANDS metadata page or the DataONE best practice guide for metadata.

What metadata should you keep?

It depends on what data you have. A good way to determine what metadata to keep is to think about what information would help someone else understand and re-use your data. To help you with this, you may consider using a metadata standard (also called a metadata schema). A metadata standard is a defined set of fields that can either be general or discipline specific. Using a standard will not only provide a rich description of your data, but also increases the likelihood of people finding your data. Check out Dublin Core, which is a commonly used general metadata standard, or find a metadata standard related to your discipline by searching the Digital Curation Centre’s disciplinary metadata directory.

How to record your metadata

Once you’ve decided what metadata you need to keep for your data, you should record this information and store it with the data. Some storage systems, like the University’s eNotebook, provide mechanisms for you to do this when you save your data. In other storage systems, like the Research Data Store and CloudStor, you may have to record your metadata manually in a README document (a text document) or a version control table.

If you have any questions or want some project specific advice, contact the Research Data team.