Data.rpi.edu Key Concepts

Printer-friendly version

Summary

data.rpi.edu aims to be a centralized access point for information about research data at RPI. It features a pilot data repository which uses the DSpace open-source institutional repository software.

The data repository has been configured to use the Handle System to generate persistent identifiers for each item in the repository.

Current goals


  • Collect and deposit at least one dataset from researchers at each RPI school, in order to be able to demonstrate the repository to researchers and institutional stakeholders.
  • Make the metadata for each dataset deposited in the repository available as linked data, by exporting and converting the metadata to triples. (We don't currently have plans to make the deposited data itself available as linked data.)
  • Make the metadata for LOGD datasets available in the data repository, with links back to the data location (or LOGD catalog page for the data). This will allow users to search for LOGD datasets through the data.rpi.edu repository. We will not deposit the LOGD data itself into the repository.

FAQ

Q: what is DSpace's capability for using URIs - to point to the actual data, as well as adding URIs for metadata (DC) terms?
A: As far as I know, DSpace's use of URIs is limited to using Handles to provide a permanent link to the record for an item. For example, see http://data.rpi.edu:8080/repository/handle/10833/4. DSpace treats the content of all of its metadata fields as literals which are indexed as text for searching. For instance, I don't believe that you can use a URI to represent a creator of an item. DSpace would just display the URI in the user interface, since there is no functionality to retrieve a related label for an entity represented by a URI. -KD

Q: will there be a triplestore associated w/ data.rpi.edu?
A: We're not sure yet what triplestore will host Alvaro's export of data.rpi.edu metadata to linked data. We'd eventually like people to be able to publish linked data to data.rpi.edu, which will require a triplestore. We're not aware of anyone wanting to deposit linked data at this point, though. -KD

Q: What Handle server are we using?
A: Handle System

Q: Is a Handle automatically generated for every item deposited in repository?
A: Yes, on the live instance (but not on the test instance, since Handle is not set up there.)