Research Data Management

File Structures & Naming Conventions

Files can be organized according to:

• Research activity (interviews, surveys, etc.)
• Data type (images, text, databases)
• Kind of material (raw data, cleaned data, published data)

Structure files in a hierarchal system, using either a deep or shallow hierarchy, using:

• The Single-Question Principle: At each level of your hierarchy, strive to make all folder names answer the same question.
• The Separation Principle: Whenever possible, limit each folder to containing only files or only other folders.
• The Domain Principle: Organize files in different domains differently

Choose a file and document naming conventions and be consistent with them

• Choose camelCase or snake_case which makes working with data files in CLIs or IDEs easier
• Consistent date formatting (e.g. YYYYDDMM)

Image source:  L. Bishop et. Al (2014) Managing and Sharing Research Data – A Guide to Good Practice;  p 69

Data Versioning

Keeping track of dataset versions makes datasets trustworthy. A given dataset may be in use by one or several data contributors or data consumers, so it is vital to make apparent which is the latest version of the data and a version history that explains changes made to prior versions.

• Simple methods such as using folder structures and naming conventions to version data
• A more powerful and collaborative tool for data versioning, is open source version control systems such as Git and DVC

There is no data community consensus on when changes to a dataset cause it to become a different dataset instead of a new version and there is no single right way to version data and track changes. Follow your discipline's best practices.

For more information on tools for version control, see our Digital Tools for Research Guide.

Data Management Plan (DMP) 

A data management plan (DMP) is a document used to describe how you plan to manage your data throughout your research project. Many funders require that a DMP be created, maintained, and submitted.
A DMP should include:

• Information about data and data format

• Metadata content and format

• Policies for access, sharing and re-use

• Long-term storage and data management

• Budget

For more information on DMPs, refer to our Data Management Plans Guide.

Metadata

Metadata provide standardized, structured information that describes a data collection in a machine-readable way. It enables data discovery, reproducibility, and preservation. A key component of metadata is a metadata schema which provides the set of standards of common components of metadata, such as identifier, creator, and date.

There are different metadata standards that you can apply to your research depending on your discipline, the type of data you are working with, and the location where you plan to publish and preserve your data. Metadata is often categorised into three categories:

• Descriptive metadata includes author, title, keywords and abstract and enable users to find resources online.

• Administrative metadata includes information about when and how a resource was created as well as file type, technical information and access rights. It includes technical metadata, rights management metadata, and preservation metadata

• Structural metadata provides information about the relationship between components in an object e.g. relating articles, issues and volumes of serial publications, or the pages and chapters of a book.

A metadata standard provides a structure to describe data with:

• Common terms for consistency between records
• Common definitions for easier interpretation
• Common language for ease of communication
• Common structure to quickly locate information
• Standards provide a uniform summary description of a dataset. 

Common Metadata Standards for Research Data

Sources: Coffey, A. M., Joy, C., Clarke, C. R., Hayes, A., McCarney, E., Madden, F., Quinn, C., Dalton, M., Stokes, D., McCabe, G., Noonan, E., & O'Dwyer, L. (2025). Navigating Open Research - A guide for early career researchers. Zenodo.

DataCite Metadata

If publishing research data to the University of Galway Community on Zenodo, DataCite is the metadata standard required for deposit into the Zenodo repository (see Guide to Publishing Research Data for more information on research data publishing).

"While DataCite’s Metadata Schema has been expanded with each new version [it is] intended to be generic to the broadest range of research datasets, rather than customized to the needs of any particular discipline. DataCite metadata primarily supports citation and discovery of data; it is not intended to supplant or replace the discipline or community specific metadata that fully describes the data, and that is vital for understanding and reuse."
 

Mandatory metadata properties

Optional and recommended metadata properties

Sources: https://datacite-metadata-schema.readthedocs.io/en/4.6/properties/overview/#mandatory-properties; https://datacite-metadata-schema.readthedocs.io/en/4.5/introduction/about-schema/

Technical Metadata

Technical metadata will be specific to your discipline and your research project. It can describe what a variable is in your dataset and how it was collected, e.g. research instrument settings, software versions, patient IDs, image resolutions, etc. These metadata fields are vital for data processing and for data reusability. When collecting data and adding metadata fields to your research data, ask yourself the following questions:

• Could someone not familiar with my data understand it?

• Does my data have a unique identifier for each data record (i.e. row of data)?

• If someone wanted to recreate my research, does my data include all the information they would need to do so?

• Do the column names in my dataset accurately describe the variables in that column?

README

A README file is a simple way to direct consumers of your data to your data documentation. A README file includes details about your data such as:

• Description of the data
• The research project it is associated with
• The version of the data and when the data was last updated
• What licensing is attributed to the dataset (for published datasets)

When creating a README file:

• Create for logical clusters of related files or data
• Write as a plain text file (.txt) avoiding proprietary file types (e.g. .docx)
• Use consistent formatting for all README files

Source: https://data.research.cornell.edu/data-management/sharing/readme/

Data Availability Statement

A data availability statement describes the datasets associated with your published work and includes a link to the relevant repository and the dataset’s persistent identifier. It can be added to the end of your published work, before the reference list. Note that a Data Availability Statement may be required for some open access publishers.  See here for examples of data availability statements.