rdap 15: virginia tech university libraries’ data service pilot with the college of natural...
TRANSCRIPT
Virginia Tech University Libraries’ Data
Service Pilot with the College of Natural
Resources and Environment (CNRE)
Natsuko Nicholls, Andi Ogier and Kyrille Goldbeck DeBose
Research Data Access & Preservation (RDAP) Summit
Minneapolis, MN
April 22, 2015
Our Mission:
To deliver a suite of research data
services to the Virginia Tech Community.
Our Vision:Build Infrastructure
Advocate for Data Management
Support Collaboration
Value Partnerships
Project Highlights
Data Profiling
Data Needs Assessment
Data Interviews
5 areas of interest
5 questions
15 faculty recruited
Q1: Data Profiles
ASCII is recommended,
but it triples or
quadruples the storage
size
Companies are taking
open formats and
creating their own
proprietary standards.
Data Profiles: Summary
● Diversity of data
○ types
○ formats
○ environments
● Data States
● Raw vs.
Summarized
● Media Issues
Common Elements: Points of Interest:
Q2. Data Workflows
Fragmented workflows are
VERY problematic.
Data Management costs
30% of project time.
Need: High-quality
workflows that allow for
creativity and spontaneity.
Data Workflows: Summary
● Lifecycles are
Complicated
● Data Management
is time consuming
● Establishing and
documenting a
workflow is time
consuming
Common Elements: Points of Interest:
Q3. Data Challenges
Need: Systems that allow
algorithms and processes
to be brought to the data.
Original data isn’t always
available.
Data Challenges: Summary
● Data challenges:
○ format
○ storage
○ versioning
○ ownership
○ copyright
○ privacy
● Workflows and
tracking data
Common Elements: Points of Interest:
Q4: Data Value-Add
Harmonization of data
users and producers.
Concern: easy to mis-
interpret some data;
analysis is dependent on
specialized knowledge.
Data Value-Add: Summary
● Value-add
○ beyond research
community
○ historical data
○ public interest
● Market impacts
● Cost of data
collection is
measurable
Common Elements: Points of Interest:
Q5: Data Management Planning
DMPs seem rigid and
limited.
Standardization hinders
innovation; different
people measure different
things.
Preserve data and prevent
loss by summarization.
Data Management: Summary
● Standards vs.
specifics
● Use DMP as a tool
to find data
● Metadata is vitally
important
Common Elements: Points of Interest:
Final Thoughts
Interview process promotes partnerships,
starts conversations.
Research Management tools are needed at a
project level, not institutional/department
level. Provide information and let project
teams make the choice.
Need additional training as to why DMPs are
important to research. Many faculty still see
them as a hindrance.
Questions?
Natsuko Nicholls
Research Data Consultant
Andi Ogier
Assistant Director
Data Curation
[email protected] Kyrille Goldbeck DeBose
College Librarian for Natural
Resources & Environment
and Animal Sciences