Single file upload: The maximum file size limit for individual file upload is 3GB. The experience for each user may vary based on the stability of the internet connection, speed of connection, file complexity, file format, and type;

For tabular data files, additional processing by the Dataverse platform means that more time is required to upload. This may result in upload failures depending on the file size and the number of columns, cases, or rows in your file. We recommend you upload files less than 500MB in size that requires this feature (see Full Dataverse User Guide)

Multiple file upload: The repository accepts .zip or .tar file bundles. It automatically unpacks your single .zip package file and presents individual files contained within to users without any hierarchy or organization maintained. Alternatively, .tar or double-zipped files can be uploaded directly, preserving your original file organization. Given the additional processing required for .zip files, we recommend you upload zip files that contain less than 500 files.

If you have many separate data files and/or metadata that you would like to upload to create or populate a Dataset or collection, consider using the SWORD API (see Advanced Dataverse User Guide), which can support batch upload of files (same size limits as described above).

It is recommended that file size be factored in when considering data reuse. Larger files can be more difficult for researchers to handle and work with. Preparing and structuring your data for reuse requires careful consideration of potential use cases, the software required for analysis, the computer system, and file disk space. The Dataverse community is working to support more disciplines and data types. We anticipate that in the future, larger datasets will be better supported and managed in the platform.

If you have questions about file size limits, please get in touch with Researcher Support.

The repository is ideal for depositing and sharing research data and supports various research data types and formats. You are welcome to upload research data, supplementary tables, and documentation. Contact the KEEP repository service to submit publications and presentations that your data supports.

We support uploading any file type, including shapefiles, images, Flexible Image Transport System files, and compressed files. However, to use the Data Explorer feature, you need to upload tabular data files (e.g., Stata, SPSS, R, Excel (xlsx), and CSV).

Some files will generate an archival file format for viewing, but users can download either the original or archival version.

See the full Dataverse User Guide for more information.

Yes. The ASU Library can work with you to copy your Harvard Dataverse to the ASU Research Data Repository.

To maintain your original statistics, we will leave the original datasets in Harvard and create new copies at ASU. A metadata record will point back to the Harvard repository. We highly recommend that ASU researchers using the Harvard interface now use ASU's installation unless there are mitigating circumstances like a journal, funder, or collaboration requirements that require Harvard's platform.

To submit a migration or questions about the process, please contact ASU Library's Researcher Support and request a repository deposit consultation.

To assist users in interpreting the data for potential replication or reuse, you should add information that helps tell others exactly how you created and used your datasets. For example, tabular data can still be hard to read if someone doesn't know what your column headers or variable names mean. Projects are varied, and you should not assume someone can figure out what you did. A readme file, code sheet, or data dictionary can help mitigate confusion and help your data align with our charge to make research data findable, accessible, discoverable, and reusable (FAIR). These elements can be uploaded directly into your dataset or added to the metadata.

For more information, see:

• How to make a Data Dictionary (Center for Open Science)

• Guide to writing "readme" style metadata (Cornell University)

• ASU Library Data Management Tutorials

Open

We recommend using CC0 and adding a suggested citation because, in most cases, data (collection of facts) might not be eligible for copyright protection and ultimately creates an unnecessary barrier to re-use. For further reading, see This post from the University of California’s Office of Scholarly Communication provides an overview of the situation and recommends the CC0 license for many kinds of data. The Open Database License (ODC-By) also has an attribution variant. See also our quick guide on licensing and data.