Data Provenance
Data Provenance is a record of where your data came from and how it reached its current form. It describes the origin of a data file, any transformations that have been made to that file, and any persons or organizations associated with that file. A data file’s provenance can aid in reproducibility and compliance with legal regulations. The Dataverse Software can help you keep track of your data’s provenance. Currently, the Dataverse Software only makes provenance information available to those who have edit permissions on your dataset, but in the future we plan to expand this feature to make provenance information available to the public.
A Dataverse installation accepts provenance information in two forms: a Provenance File or a free-text Provenance Description. You can attach this provenance information to your data files in a Dataverse installation as part of the file upload process, by clicking Edit File -> Provenance:
This will open a window where you can add your Provenance File and/or Provenance Description:
A Provenance File is the preferred way of submitting provenance information to a Dataverse installation because it provides a detailed and trustworthy record. Provenance files are typically generated during the process of data analysis, using provenance capture tools like provR, RDataTracker, NoWorkFlow, recordr, or CamFlow.
Once you upload a provenance file, the Dataverse installation will need some additional information in order to accurately connect it to your data file. Once provenance file upload finishes, an input box labeled “Connect entity” will appear under the file. Provenance files contain a list of “entities”, which include your data file as well as any objects associated with it (e.g. a chart, a spellchecker, etc.). You will need to tell the Dataverse installation which entity within the provenance file represents your data file. You may type the name of the entity into the box, or click the arrow next to the box and select the entity from a list of all entities in the provenance file.
For more information on entities and the contents of provenance files, see the W3C PROV Model Primer.
Once you’ve uploaded your Provenance File and connected the proper entity, you can hit the Preview button to view the raw JSON of the Provenance File. This can help you confirm that you’ve uploaded the right file. Be sure to double-check it, because the Provenance File will made permanent once it’s finalized. At that point you will not be able to replace, remove, or otherwise edit the Provenance File. This ensures that the Provenance File maintains a stable, immutable record of the data file’s history. This finalization of the Provenance File happens at different points depending on the status of your data file. If this is a brand new data file that has never been published before, then its associated Provenance File will be made permanent once you publish the dataset. If this data file has been published in a previous version of your dataset, then its associated Provenance File will be made permanent as soon as you upload the Provenance File and click “Save Changes” on the warning popup.
A Provenance Description allows you to add more provenance information in addition to or in place of a provenance file. This is a free-text field that allows you to enter any information you feel might be relevant to those interested in learning about the provenance of your data. This might be a good place to describe provenance factors like what operating system you used when working with the data file, what functions or libraries you used, how data was merged into the file, what version of the file you used, etc. The Provenance Description is not as useful or trustworthy as a provenance file, but it can still provide value. Unlike the Provenance File, the Provenance Description is never made permanent: you can always edit, remove, or replace it at any time.
You can return to attach provenance to your data file later on by clicking the “Add + Edit Metadata” button on the file page, and then clicking the “Edit File -> Provenance” button.