Research Data Management

Without description, data is hard to understand and use. Make your data FAIR (findable, accessible, interoperable, reusable) by describing it with metadata (data about data). Metadata is sometimes captured through deposit in data repositories, but you can also prepare data dictionaries, codebooks and README files to further describe and contextualize your work. 

README files are plain text documents that sit at the top level of project folders and describe the purpose of the project, contact details, and organization of files. Including a README with your work helps ensure that future users will understand the data, any terms, and more. 

README files should include: 

• Title
• Principle Investigator(s)
• Dates/Locations of data collection
• Keywords
• Language
• Funding
• Descriptions of every folder, file, format, data collection method, instruments, etc. 
• Definitions
• People involved
• Recommended citation

For more information about writing README files for data, we recommend Cornell University's comprehensive guide to readme files.  

Metadata standards or schemas consist of specific elements used to describe or document your data. It helps make your data easier to share and publicize their data and also locate and retrieve data sets.

The three most common categories on metadata are:

1. Descriptive: describes the content and context of your data and both the dataset and item level. Example: title, author, keywords
2. Administrative: includes information needed to use the data. Examples: software requirements, copyright, licensing.
3. Structural: describes how different datasets relate to one another, or any processing or formatting steps that were undertaken. Example: Information about the relationship between datasets in a database, file formats. 

Adapted from Data Management Plans: Documentation and Metadata by UBC Library Research Commons under Creative Commons Attribution 4.0 International. " 

Some disciplines have established metadata standards. In addition, some data repositories have their own standards. There are also several general purpose schemas that you can adapt to fit your needs. To find a metadata standard, the Digital Curation Centre (DCC) has an inventory of discipline-specific metadata standards.

File names and a simple hierarchy will make files easier to locate. Set up conventions for your project, document them for all team members and be consistent! 

Recommendations:

• Use dates in YYYYMMDD format (20220403)
• Use a short identifier (e.g Project Name or Grant #)
• Include a summary of content (e.g Questionnaire or GrantProposal) as file name
• Use_as delimiters. Avoid: &,*%#*()!@${}[]?<>-
• Keep track of document versions either sequentially or within a unique date and time
• Make folder hierarchies as simple as possible

Creative Commons Attribution 4.0 (CC BY)

Adapted from the University of British Columbia's "Organize"

Directory structures refer to the organization of folders and files. In creating a directory structure, you want to find a balance between making your folder structure too deep or too shallow. A directory hierarchy might look something like this:

• Root directory (top level folder)
  • Subject directories (subfolders)
    • Files

To set yourself up for success, document the rules underlying your structure's organization. It will be easier to file your data correctly if the rules can be referenced. 

Further resources can be found here:

• UBC's Directory Structures Guide for guidance and examples.
• Chapter 3 in: Briney, K. (2023). The Research Data Management Workbook. Caltech Library. https://doi.org/10.7907/z6czh-7zx60 This has a fun and practice exercise that you can do to help you create a file organization system for a research project.