Embed Size (px)
Citation preview
Embedding research datamanagement behavioural changewithin policies, systems and human support infrastructuresNewcastle University1 Niall O’Loughlin, Lindsay Wood, Megan Quentin-Baxter, Phil Heslop, Simon Kometa and Janet Wheeler
Introduction• Management of research data has significant
academic and financial implications for theUniversity. Direct research income was £88 millionin 2010-2011 and REF associated Quality-Relatedfunding returns £35 million annually
• Public funders are now mandating that their researchdata are available for verification and re-use
• The University is addressing these challengesthrough the JISC-funded 18 month cross-department iridium project, due for completion inMarch 2013
• The initial audit of current research datamanagement status and staff attitudes is complete.The next stage is to incorporate these into outputsi.e. policies, research systems, and trainingtogether with final report recommendations and acosted business case
MethodsRDM requirements gathering: An online survey wasdistributed to all active Principal Investigators (932 staff).Representative academics and research staff across theinstitution were invited for a face-to-face interview (163staff). Transcription and thematic analysis were carried out.2
RDM policy framework development: A review ofexternal/internal policies and guidance documentation wasconducted. This was mapped to the Digital Curation CentreRDM lifecycle model.3
RDM systems development: Current research businesssystems were reviewed to assess their suitability andcapability as data management and catalogue systems.Required functionality was documented.
ResultsRDM requirements gathering: The online survey wascompleted by 128 research projects, representing 15% ofactive academic staff. The key findings were that:
• 31% of projects’ data storage needs were satisfied bythe institutional free allowance (4GB)
• 64% of projects’ data location was serviced by theinstitution
• 51% of projects were happy to release their publication-associated data within 1 year
• 23% of projects had a formal data management plan
Additional findings from project responses are reported inFigure 1a-d. Face-to-face interviews have been completedfor 27 staff to date with emerging themes reportedincluding:
• clarifying RDM institutional expectations and trainingopportunities, future policy/systems integration withexisting structures, archiving guidance and datacollaboration tools.
RDM policy framework development: The policy analysisresulted in two outputs; 11 general high level principlesconstituting the Research Data Management Policy andsupported by the Code of Good Practice. This broughttogether all relevant institutional guidance into oneaccessible document.
RDM systems development: High quality project, personand publication data existed in the University’s currentsystems, however none dealt specifically with researchdata. In response, a data catalogue system wasdeveloped; this used existing collected data to provide richderived metadata (Figure 2), thereby minimising workloadduplication and increasing discoverability and the likelihoodof re-use. The metadata records can be exposed (internallyor externally), as a human readable website and/or in aresearch business system machine readable format.
Discussion• Moderate storage quota increases would allow
most projects data needs to be satisfied
• Data retention up to 10 years was most commonand is in line with funder requirements
• Institutional storage locations (with definedoperational service standards) were used, in part,by most (i.e. two-thirds) of projects
• The Research Data Catalogue provides ametadata index (CERIF compliant4) of institutionalresearch data, in lieu of an institutional repository
• Outputs will be trialled with exemplar projects andevaluation undertaken
MyProjects
Projects
People
MyImpact
PeoplePublications
Research Data Catalogue
Machine readable formats
(i.e. XML/JSON)
Human readable website
(%
)0
70 (a)
(%
))
070 (b)
(%
)0
70 (c)
(%
)0
70 (d)
Proj
ects
40
5060
Proj
ects
40
5060
Proj
ects
40
5060
Proj
ects
40
5060
2030
2030
2030
2030
<
010
< 4 GB
>
16 GB <
<
6
4 GB < >
1
<
5 T
0
0.5 TB
TB
1 T
64 GB <
1 TB < >
B <
10 <
1
100
> 1 P
>
PB
1 < > 1
00 < >
01
PB
0
InI Ac
AAcade
roPrnstituti
Pr
rct
oject dem
P OthO
theO�O -c�-cam
ExPersona er
O xtExtet rn
er rn
xtExteterrnal l
oAt
nal
At homl
ImIm
010
di
1 m
>
6 mo <
< >
2 yrs
mmedi
2
1 m < >
<
< >
5 + yrs
s
Re rs < >
Rre
etirem
DeD ateath ve
Never
OthOtherth
< 1 yr
010
1 < 5 yrs
5 < 10 10 y
10 < 25
25 + + yrs
B
>16 GB
<
< >
Da
< >
ta st
> 64 G
< >
5 T
0.5 TB
GB
1 T
> 1 T
> 0.5
>
st
TB
< > 1
1 T
orage spac
0 T
10 T
0 T
100 TB
TB
> 0
0 T
> 1 PB
c
B
10
e Data st
onal nit
t m mtio mic Uni
torage locat
ysmpu ys
y man
nal sy
stestem
mnaged
tion
s
u
al rv
pusl s
ce
al cloud vice
mel ser
atdiatetely
> 6 mo
<
< > 1 y
di
Da
>
o
t
1 yr
> 5 yrs
ntment
ta release
rs
Data re
yrs
5 yrs
etention
rs
Iridium foil image cc: by-sa Dschwenhttp://commons.wikimedia.org/wiki/User:Dschwen
T: +44 191 222 5499W: research.ncl.ac.uk/iridium/Blog: iridiummrd.wordpress.comTwitter: @iridium_mrdEmail: [email protected]
Niall O’Loughlin,Policy and Information OfficerResearch & Enterprise ServicesNewcastle UniversityNewcastle upon Tyne NE1 7RU
Figure 2. Schematic diagram of dataflows to and from the research
metadata catalogue system
References1 University Research Office, the Digital Institute,
the University Library, Information Systems & Services and MEDEV,School of Medical Sciences Education Development (research.ncl.ac.uk/iridium)
2 Braun, V. and Clarke, V. (2006) 'Using thematic analysis in psychology',Qualitative Research in Psychology, 3: 2, 77-101;
3 www.dcc.ac.uk/resources/curation-lifecycle-model (accessed June 2012);4 www.eurocris.org/Index.php?page=featuresCERIF&t=1 (accessed June 2012).
Figure 1. Bar chart plots of online survey percentage of projects reporting space, location, retention and release requirements
Iridium poster June2012_Layout 1 20/06/2012 16:30 Page 1
