Thursday, 31 July 2008

Digital Repository Services for Research Data Management

More notes from the Edinburgh Repository Fringe:

Luis Martinez Uribe discussed the findings of the Oxford Research Data Management Project - Scoping digital repository services for research data management (heard here for the first time!).

He explained the relative complexity of Oxford University's devolved infrastrucure and the Oxford Federation of Digital Repositories containing a whole range of digital objects and collections

The overall aim of the project was to scope requirements for repository services to manage and curate research data i.e. using RIN classifications of observational; experimental; derived; simulations; reference data

He conducted scoping study interviews in order to establish how researchers use their data - interviewing 37 researchers across disciplines.

His findings looked at 4 key areas:

Funding: he found that there was no detailed planning in place in funding bids, funding came from a variety of sources, any plans that did exist were fairly generic, although researchers did indicate an awareness of need to make data available although no retention plans etc were in place.

Data collection: Luis reported the varied origins of research data, format and file size variations,indeed much data collected was collated from printed sources

Processing: the majority of research data was stored on departmental servers or desktops, no/little metadata, poor annotation, sharing by email/portable media, with ensuing storage and sharing problems particularly with big datasets

Publication - he established that there were few deposits in national archives (time/effort on the part of researchers), some data was published on the web, and there was general agreement on the usefulness of linking data and publications

One of the main issues that emerged from the exercise was SUPPORT, or rather the lack of it. No mention of the role of librarians with any support that did exist coming directly from departmental IT officers. It was also evident that researchers were very much in favour of a 'federated' approach rather than centralised models with the impositions that could/would be entailed.

As a result of the scoping study Luis identified that there was a need for secure and friendly solutions to store and share data which can only be achieved by concerted thinking at policy level about a sustainable infrastructure(s) to publish and preserve research data. However to reiterate, he also identified that a major requirement is support, support, support - support for data management plans, data formats, data sharing policies and guidelines, storage and curation etc.

Luis sees the next steps being to consult more with service providers, libraries, information and computing services departments in order to address and validate requirements - He also plans to implement the DAF (Data Audit Framework) methodology currently in development by HATII/DCC at the University of Glasgow to audit oxford's data resources in conjunction with DISC-UK DataShare.

Perhaps much of what was said is already in many of the minds of practitioners however this first public airing of these findings offers a degree of clarity and articulation on the issues surrounding institutional data auditing.

I'm hoping to catch the end of Prof. Dave de Roure's (Southampton University) presentation on 'How repositories can avoid failing like the Grid' - if I don't get round to blogging this all the main session 's have been recorded and will be available at a comfortable desktop Edinburgh Repository Fringe cinema near you!

Stuart Macdonald
DISC-UK DataShare

No comments: