Dataverse running locally

[Charbel Zeinaty]

Hello,

I am on a internship working at an institution that want an application for their data management plan. I have been advised to work with dataverse. However my question is, can I install the application internally on the software and use it on the enterprise intranet without the need to publish any data or metadata?

I will highly appreciate any feedback or comment,

[danny...@g.harvard.edu]

Hi, it's possible to run Dataverse locally, but since several core features of the application are around publishing metadata and data you may run into some challenges. For example, during the publishing process Dataverse requires that a Persistent Identifier (PID) be created for the dataset, with the intent that the PID be publicly discoverable. You could perhaps investigate some of the settings in the Developer Guide regarding the FAKE DOI provider (http://guides.dataverse.org/en/latest/developers/dev-environment.html#configure-your-development-environment-for-publishing), but again, there may be other areas that you run into a feature mismatch.

Hope this helps,

Danny

[Charbel Zeinaty]

Thanks Danny for your reply,

Yes I know about the DOI and the fake one but I was scared about other mismatches as like you just wrote.

The thing is that I am asked to work with dataverse as an application to exchange information locally without taking full advantage of the software potential. Can I publish datasets and dataverses locally on intranet server does this work ? is it possible in the future to migrate again to the full dataverse potential, what are the risks (data loss ?)

Best Regards,

[danny...@g.harvard.edu]

Hi, it's hard to say. Others in the community may have experiences making a local Dataverse installation public later, and may be able to share experiences. I think the data itself would be fine, but the PIDs are the bigger issue, and if you wanted to make it public later there could be issues with having those PIDs still function. Hopefully others in the community will have more information. 

What's the reason that the metadata and data can't be shared publicly? 

[Philip Durbin]

If you look at the crowdsourced spreadsheet* of Dataverse installations you'll find that a few of them use the "authority" 10.5072 which is only supposed to be used for testing. (I assume they're using the FAKE DOI provider as well.) These installations seem to work fine (more on this below) but in a perfect world, all installations would get their own DOI authority or use Handles. As Danny said, data citation is very important to this community.

Here are the features I believe you lose if you don't configure Dataverse to use proper persistent identifiers (DOIs or Handles):

- The ability for the data to be cited.

- The ability to click the URL form of the persistent identifier (PID) in the citation of the dataset have it "resolve" back to the dataset. This is important when you share the citation to the dataset, of course.

Using DOIs instead of Handles gives other features:

- When a dataset is published, metadata is sent to DataCite's search index.

- Citations to datasets can be retrieved via Make Data Count.

Since you mentioned migration, there are APIs in Dataverse to migrate from Handles to DOIs but I'm not aware of any tools to migrate from fake DOIs to real DOIs.

I can't think of any risks (you mentioned data loss), apart from the features you miss out on above and that you might hear again and again that you should be using real persistent identifiers for your data. :)

I hope this helps,

Phil

* Linked from https://github.com/IQSS/dataverse-installations

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/35e106ff-b008-49aa-a123-9034d829f912n%40googlegroups.com.

--

Philip Durbin
Software Developer for http://dataverse.org
http://www.iq.harvard.edu/people/philip-durbin

[Charbel Zeinaty]

Hi Danny,

Thanks for your reply. THe only reason is that at the moment no budget is available for a Datacite account, but they want an application that would help store and manage and search for their own data at the moment.

Regards,

[Charbel Zeinaty]

Hi Philip,

I thanks you too for your reply, I think I got a clearer idea on what is coming and how to proceed now.

Thank you.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/35e106ff-b008-49aa-a123-9034d829f912n%40googlegroups.com.