Working with data

We all want to keep our data safe and secure, but there are several aspects to this which you should consider to have confidence that your data is as safe as it can be.

Preventing loss and corruption

Files can be lost accidentally in many different ways. Even if they are not lost completely, they can occasionally become corrupted. If a file is severely corrupted it may be unusable, but even subtle corruption may introduce errors which go unnoticed while affecting the outcome of your research.

Backing-up data at University of Westminster

See Store files (intranet link, login to access) for further information including a link to the University's back up policy and process.

Things to think about

  • Regular backups: (ideally automated) to several different locations will ensure that if one copy is lost or corrupt, you can easily get it back. When deciding how often to back up, think about the maximum number of days' work you would be prepared to lose.
  • Checksum tools: A checksum is a file's digital signature, which can then be used to detect unexpected changes in their contents.
  • Non-digital data: If you have data which are not kept on a computer, you should make sure they are protected too.

Controlling access

Preventing unauthorised access

In many cases, you may wish to restrict access to your data to a specific list of individuals. This might be because it is commercially sensitive to you or an industrial partner, or includes sensitive personal information covered by the Data Protection Act.

If you are collecting or using research data about individuals, you should read the University's Data Protection policy, which includes information about academic research.

Things to think about

  • Legal requirements: You may be under legal and/or contractual obligations to protect your data. If you're not sure, you can discuss this with Research Development Managers within the university's Commercial and Business Development Support team (intranet link login to access), who can give you advice on Intellectual Property, your collaboration or consortium agreements. See also Research grants and contracts financial regulations (intranet link login to access). The Information Security and Compliance team can provide guidance and support relating to the Data Protection Act and Freedom of Information.
  • Use of secure systems: One way to restrict access is to use a password-protected system. Commercial services such as Dropbox may be convenient, but are unlikely to provide sufficient protection against unauthorised access. See store files (intranet link login to access) for further information about how to classify your data and which storage options are appropriate.
  • Secure passwords: Passwords are often the weak link in any secure system. Make sure you choose passwords that are long and difficult to guess. Writing them down is OK, as long as you protect your written-down password very well, just like you would with your house or car keys. See strong passwords for further guidance.
  • Encryption: You will sometimes need to send data to people who don't have access to your secure storage system. Encrypting a file before you send it via insecure means (e.g. email) ensures that the contents can only be read by someone who has the key.

Ensuring usability

An often-overlooked aspect of data safety is ensuring that it remains usable. Students and staff arrive and leave on a regular basis, and often it can seem easier to repeat a whole set of expensive experiments rather than try to understand data left behind by researchers who have left the university.

Things to think about

  • Documenting data: Record information about the structure and format of your data and the process you went through to obtain it. In some cases this can be stored in the data files themselves; if not, it can be stored in a "read me" document in the same folder as the data.
  • Using standards: Be aware of standard file formats and standard nomenclature (such as letters used for variables) used in your field. Consider using files in open formats so that they can be read by a variety of software.