What is a data repository?

A data repository is a service that commits to long-term storage and preservation of data, provides a persistent identifier (Digital Object Identifier, or DOI) that points to the data and a description of the data, and typically makes data publicly accessible (sometimes with restrictions).

Can't I just email people my data when they ask for it?

Funders increasingly require data to be deposited in a repository for long-term preservation and sharing rather than simply being shared "upon request", since the latter model results in low levels of data sharing and lost data. Submitting to a data repository decreases the burden of sharing for both parties.

Choosing a data repository

Data repositories range from hyper-specific (accepting only one type of data and requiring very specific metadata, such as GenBank for DNA sequences) to general (accepting any type of data from anyone on any topic with minimal required description). The best choice for your data is the place where it is most likely to be found by others who are generating or interested in similar data.

To select a repository:

1. Determine whether your funder or publisher specifies a repository that you must use.
2. If not, search for a repository that is specific to your data type using the FairSharing repository database.
3. If none exists, search for a repository that is specific to your discipline/research subject using the FairSharing repository database.
4. If none exists, select a generalist repository from this list provided by the National Libraries of Medicine. We recommend Figshare, since Wesleyan has an institutional subscription that allows us to provide personalized control of permissions and storage as well as robust support. You can instantly create an account on Figshare simply by signing in using your SSO credentials and easily submit your data and obtain a Digital Object Identifier.

Contact a librarian using the information in the sidebar for assistance with selecting a repository and using Figshare.

If you've been following your data management plan, your data should already be well-organized and nearly ready to publish. The next steps are to:

1. Decide what data you will share. It may be tempting to share only your final figures, or to share every single piece of data that was collected. However, you should consider what data will be needed by another researcher (or your future self) in order to understand and reproduce your research, as well as legal or ethical requirements.

   • De-identifying data: If your data contain personal information, you will need to share it according to your IRB-approved protocol in a way that protects individual research participants. This guide from Johns Hopkins Libraries provides basic and advanced techniques for data de-identification.

2. Organize your data. Name files uniquely and in a way that logically reflects your research, and organize them in a folder hierarchy. Files over ~5GB in size may cause problems when being uploaded to or downloaded from a repository, so if possible try to keep files smaller.

3. Document your data. Ensure that code is well formatted and commented. Name columns in tabular data descriptively and with units. Prepare a plain text or markdown format readme file that includes information about the research project, file naming conventions, descriptions of each data file, meanings and units of each column (for tabular data), and any other information needed so that your data could be easily understood and interpreted by someone familiar with your field of research. You can download a template readme file here. Edit it in your default text editing software to reflect your own project and save it with "readme" in the file name (e.g. "readme.txt").
   • Some sources recommend creating a separate readme for every data file. This depends on the nature of the files: how similar are they to one another? It's up to you to decide whether your data will benefit most from one readme file included at the top level of your folder hierarchy, a readme for every folder, or a readme for every file.
4. Publish! If you choose to share your data on Wesleyan's Figshare data repository, you can sign in using SSO, deposit your data, and obtain a DOI within minutes. Contact the library for help with Figshare. If a different repository is more appropriate for your data, we will also do our best to assist you with your deposit.

Additional resources on preparing data for publication:

• Preparing data for deposit, Northeastern University
• Preparing tabular data for description and archiving, Cornell University

Have questions about preparing and publishing your data? Know of a resource you think we should add? Let us know.