Attitudes and aspirations in a diverse world

1
Attitudes and aspirations in a diverse world

• The Project StORe perspective on scientific
  repositories
• Graham Pryor 22nd November 2006

Digital Data Curation in Practice - 2nd
International Digital Curation Conference,
Glasgow 21-22 November 2006
2
StORe Guide

• Whats in StORe?
• Curation and preservation issues
• Attitudes and aspirations
• Research data
• Repositories
• Metadata
• Ownership and support
• Too huge to handle?

3
(No Transcript)
4
Digital Data Curation

• Definition the actions needed to maintain
  digital data and other digital materials over
  their entire life-cycle and indefinitely for
  current and future generations of users. These
  actions include not only the processes of digital
  archiving and preservation but also
• all of the processes that are essential to good
  data creation and management, as well as the
  capacity to add value to data to generate new
  sources of information and knowledge.

5
Whats in StORe? Aims 1

• Attach new value to the intellectual products of
  academic research by providing two-way links
  between source and output repositories

6
Whats in StORe Aims 2

• Surveys to identify workflows and norms, problems
  and desirable enhancements to source/output
  repositories
• A generic technical specification for functional
  enhancements to source and output repositories
• Pilot middleware that demonstrates a
  bi-directional link
• Independent evaluations of the pilot middleware
  and recommendations for future development as a
  generic platform for linking repositories

7
Whats in StORe? Survey

• Genuine desire to contribute to the wealth of
  knowledge
• Awareness of the critical need to assign and
  maintain appropriate metadata

• Dual deposit of data and publications already an
  accepted concept
• International strategies for data deposit and
  data preservation

8
Whats in StORe? Survey

• Cultural and organisational barriers to deposit
  of research data in repositories
• Inherent culture of self-sufficiency in the
  generation and organisation of data

• Limited inclination towards voluntary deposit in
  open access source repositories
• Institutional output repositories not on the
  agenda of most researchers

9
Research Data

• Features of source data
• Often large and complex
• Can be impenetrable without local tools
• May seem ambiguous to project outsiders
• Are frequently held on standalone equipment
• Commonly comprise several data formats

• From the StORe survey
• Physics raw data sets as large as petabytes
  (1015 bytes) may be generated or analysed using
  software written within the project
• Biosciences need to describe how data were
  produced, the laboratory conditions and
  methodology
• 70 of bioscience source data are not networked
• Chemists data stored in numerous sub-folders
  (spectra, images, etc.) describing one process

10
Research Data

• ..it would have to be everything associated
  with that compound. There is no point having an
  NMR without a picture of what it is. Then its
  useful to have a synthesis scenario and say oh
  that could fit with that but I want proof and
  then that really is a paper. You know you can
  waste a lot of time trying to follow what people
  have done before that isnt properly published
  and never have worth. Its not always, but is it
  worth the risk of wasting too much of your time?

• Chemistry
• data sets
• links
• between
• complex
• clusters

11
Research Data

• Physics data types

12
Research Data

• Archaeology
• file types

13
Repositories

• Source repository development is discipline-led
• Large number of established services
• - we suggested Archaeology Data Service,
  Brookhaven National Laboratories, CERN, GenBank,
  National Crystallography Service, NERC Data
  Centres, Protein Structures Database,
  SuperCOSMOS, UK Data Archive, UniProt
• - to which were added 99 others
• Some international strategies/
• - Astronomy (Virtual Observatory)

14
Repositories

• Source repository development is discipline-led
• Large number of established services
• - we suggested Archaeology Data Service,
  Brookhaven National Laboratories, CERN, GenBank,
  National Crystallography Service, NERC Data
  Centres, Protein Structures Database,
  SuperCOSMOS, UK Data Archive, UniProt
• - to which were added 99 others
• Some international strategies/mandates/
• - Astronomy (Virtual Observatory)
• - Biosciences (sequence data)

15
Repositories

• Source repository development is discipline-led
• Large number of established services
• - we suggested Archaeology Data Service,
  Brookhaven National Laboratories, CERN, GenBank,
  National Crystallography Service, NERC Data
  Centres, Protein Structures Database,
  SuperCOSMOS, UK Data Archive, UniProt
• - to which were added 99 others
• Some international strategies/mandates/dual
  deposit
• - Astronomy (Virtual Observatory)
• - Biosciences (sequence data)
• - Chemistry (Crystallographic Data Centre)

16
Repositories

• Low awareness
• of repositories

• 65 of the chemists surveyed had not used a
  repository and were not familiar with the idea of
  open access repositories

17
Repositories

• Low awareness
• of repositories
• Low volume of repository use

• 65 of the chemists surveyed had not used a
  repository and were not familiar with the idea of
  open access repositories
• Many social scientists did not associate
  repositories with their research agenda

18
Repositories

• Low awareness
• of repositories
• Low volume of repository use

• 65 of the chemists surveyed had not used a
  repository and were not familiar with the idea of
  open access repositories
• Many social scientists did not associate
  repositories with their research agenda
• Repositories are only one of many potential data
  sources/archives used by researchers

19
Repositories

• Low awareness of repositories
• Low volume of repository use
• Low rate of source data deposit

20
Repositories

• Low awareness of repositories
• Low volume of repository use
• Low rate of source data deposit
• Output repositories
• prefer publisher over institutional
• prefer Google type searching

21
Metadata

• All disciplines an awareness of the importance
  of appropriate metadata
• Improvements to source repositories? Better
  metadata ranked highest
• Metadata assignment considered challenging
  intellectually and in the demands on ones time
• Yet
• Evidence of lack of standard structures
• Metadata assignment often almost an afterthought
• One third of StORe respondents believed no
  metadata were being assigned

22
Metadata assignment
23
Metadata

• Where researchers are familiar with metadata they
  possess an in-depth knowledge of its use,
  applications and functions
• The assignment of metadata automatically (or by a
  process that relieves the depositor of doing it)
  is preferred
• Quote from theoretical chemistry interview
• Well, theres lots of different types of
  metadata. There is metadata for discovery, there
  is metadata for semantics, there is metadata for
  intellectual property and so on and so forth.
  They are all important. If I find some piece of
  information and its not on open access then I
  cant use it. If I find some piece of metadata
  and its in a language that my machine does not
  understand and there is no metadata, then it is
  uninterpretable, I cannot use it. If I am
  particularly concerned about the quality of data
  I need provenance metadata. So there are
  different needs for different people...

24
Metadata

• Need for improved and universal standards
  acknowledged
• A clear link identified between the condition of
  metadata used and the level of support from
  information specialists
• Recognition of the need for different metadata
  for different phases of research lifecycle (raw,
  processed, published data and beyond) and to
  assist cross-discipline interpretation

25
Ownership Support

• Working culture self-reliance and a constant
  pressure to deliver
• Qualified enthusiasm for deposit in source
  repositories producer or consumer
• Anxiety over predatory access and IPR
• Storage methods protectionism?
• Provision of specialist support less a case of
  unavailability as not sought

26
Too Huge to Handle?

• One of the aspects that the Chemistry
  interviewees commented upon was that there should
  be a wider organisational/institutional
  requirement that supports and manages the
  repositories, should they be source, output or
  institutional.
• sustainability depends on a business model.
  And its a major problem that confronts everybody
  at the moment in aggregating data, whether it be
  raw data, processed data, metadata, primary
  publications, abstracts, things like that

27
Too Huge to Handle?

• Embedding of data management expertise within
  domains
• Expensive?
• Interventionist?
• Too large and too difficult?

28
Too Huge to Handle?

• Embedding of data management expertise within
  domains
• Expensive?
• Interventionist?
• Too large and too difficult?
• Poor investment decisions can have major
  implications on how much information can be
  preserved, and how effectively
• Chris Rusbridge, http//www.ariadne.ac.uk/issue
  46/

29
Too Huge to Handle?

• MRC 1 million data sharing and preservation
  initiative
• - http//www.mrc.ac.uk/strategy-data_sharing.htm
• - initial focus on 4 to 6 unique datasets of
  long term value
• - engage community support longer term business
  plan
• Virtual Observatory
• Exploit information management and curation
  experience in the university libraries and build
  on long-term institutional commitments to
  preservation Bob Hanisch
• http//www.arl.org/sparc/meetings/ala06/HanischPP
  T.pdf

30
END
http//jiscstore.jot.com/SurveyPhase