Managing Open Scientific Data: 
From FEAR to FREEDOM 

Data, 
computers 
and 
collaboration 
extending the 
frontiers of 
science
Key differences between open publications and 
open data
Publications
 Available in required 
formats at the time of 
publication or earlier, 
no additonal effort 
required
 Incentives to publish 
and increase impact.
 Libranians play a key 
role.
 One size fits all.
Data
Require curation and 
substantial inputs from 
researchers.
No incentives to curate 
and share data.
Research managers and 
data scientists to play a 
key role.
Always contextual
Why OPEN DATA?
“Ten or 20 years ago we might 
have been able to repeat an 
experiment. They were 
simpler, cheaper and on a 
smaller scale. Today that is not 
the case.
So if we need to re-evaluate 
the data we collect to test a 
new theory, or adjust it to a 
new development, we are 
going to have to be able reuse 
it. That means we are going to 
need to save it as open data.“
Heuer, Former CERN Director
Research Data Management
The process of 
organising, 
manipulating, 
storing, curating and 
using research data 
to enhance its 
preservation and 
access into the 
future.  
Credit: Michael Gibbs
THE PROMISE of RDM
Digital curation of data for public 
release, that is making scientific 
data available and useful to 
unknown audiences, and for 
unanticipated purposes.
RDM REQUIRES COLLABORATION AND 
TRUST.
FEAR of CHANGE
“Individuals, work teams, departmental 
units, leadership groups and even whole 
organisations are naturally averse to 
change as they unknowingly protect 
themselves from making the very changes 
they most desire.”
Robert Kegan and Lisa Laskow Lahey, Immunity to Change
Managing data? What DATA? 
DATA is 'everything', 
means different things to 
different people – from 
BIG data to LITTLE data, 
RAW to PROCESSED data.
Open data is free for 
anyone to use, reuse and 
redistribute it – subject 
only and at most, to the 
requirement to attribute 
or share-alike.
Emergent agreement: Research data are 
the evidence used to support the findings 
and arguments in science, research or 
scholarship. 
 data captured from instruments
 derived data
 Documents
 spreadsheets & databases
 lab notebooks ???
 visualisations
 models, 
 software
 images, measurements and numbers.
What is 'the evidence'? 
What is 'data'?
 DATA is always contextual – the definition varies not 
only across scientific disciplines but also across 
different projects and different people.  
 MINIMUM STANDARD:
 'Data'
 'Metadata'
 'Software'???? – IP clearance and version control 
are necessary. If software cannot be shared, the 
version used should be listed. 
What is 'data'?
 DATA is always contextual – the definition varies not 
only across scientific disciplines but also across 
different projects and different people.  
 MINIMUM STANDARD:
 'Data'
 'Metadata'
 'Software'???? – IP clearance and version control 
are necessary. If software cannot be shared, the 
version used should be listed. 
So what is 'data'?
 DATA is always contextual – the definition varies not 
only across scientific disciplines but also across 
different projects and different people.  
 MINIMUM STANDARD – OPEN DATA:
 'Data'
 'Metadata'
 'Software'???? – IP clearance and version control 
are necessary. If software cannot be shared, the 
version used should be listed. 
METADATA 
Not just 'data about data'.
METADATA explains, describes, locates or 
otherwise makes it easier to retrieve, use, 
or manage the primary data. 
National Information Standards Organisation
MAKES DATA SENSIBLE and USABLE.  
ISSUE: No established protocols for 
recording metadata.
Open 'by default' or 'controlled access' 
DEFAULT
 Data underpinning 
scientific publications
 Most geospatial data to 
facilitate access to 
public infrastructure 
and manage public 
health issues
 Genomic data
CONTROLLED
Most research data where 
data interpretation, 
assumptions and 
detailed analyses are 
required, e. g. particle 
physics, clinical trials.  
LEVELS OF CONTROL (CERN)  
Level 1: the content of 
scientific publications 
and data, such as figures 
embedded in these 
publications. OPEN 
Level 2: simplified data 
formats for immediate 
re-use, for example, in 
education outreach 
activities. OPEN  
Level 4: raw data and 
software that enable full 
reconstruction of the 
CMS experiments. 
CLOSED
Level 3: simulation data, 
along with software, 
workflow analyses and 
other documentation 
needed to reproduce 
published results.CLOSED 
Barriers to sharing
Too many stakeholders, too many interests
 Data collectors/authors
 Data collaborators and 
colleagues
 Software developers
 Data curators
 Data managers
 Research team
 Principal Investigator
 Research subjects
 Citizen scientists
Research funders
Policy makers
Management of research 
organisations
Librarians
Data custodians
Publishers
FEARS of researchers
No time to curate data.
Validity of secondary 
data analyses
The lack of incentives to 
curate and share data
The lack of incentives to 
(re)use somebody else's 
data.
FEARS of policy makers
 Large investments in data, resources, and 
knowledge infrastructures.
PRESENT COST vs FUTURE BENEFIT
 Safeguarding the value of open data.
– The capacity to use is limited to 
elites. 
 Protecting commercial interest and IP.
 Protecting privacy of research subjects.
Where are the Open Data Elites?
Row 1 Row 2 Row 3 Row 4
0
2
4
6
8
10
12
Column 1
Column 2
Column 3
KEY LESSONS LEARNT
The value of open scientific data lies 
primarily in their quality and the potential 
for future use and re-use.
More open scientific data does not 
necessarily mean more science, 
reproducible science or data-driven science.
Open Data requires an Open Mindset.
Key observations
 Open scientific data requires strong leadership and 
change management skills within research 
organisations. 
 Unclear meaning of 'research data' acts as a barrier to 
sharing. 
 Incentives for researchers and their organisations to 
curate, share and re-use data are required.
 Resources are finite and choices must be made about 
what data to curate and maintain. Need to develop 
'data packages/products' rather than 'open data'.
 The value of open scientific data lies in their quality 
and potential for future (re)use. Metadata are the key 
to improved quality.