Jonathan Crabtree

1
Vision for Sociometric Analysis of Long-tail
Science Data
Jonathan Crabtree Assistant Director of Computing
and Archival Research UNC Chapel Hill, Odum
Institute
2
(No Transcript)
3
Motivation
4
How do we cope?

• Metadata creation
• Data discovery
• Data organization

http//blog.whatarmy.com/overwhelmed-by-social-med
ia/
5
Data Similarity

• What can we learn by comparing?

http//www.gotyourhandsfull.com/2009/09/twins-and-
more-on-a-tuesday-post-natal-depression-in-multipl
e-birth-mums.html
6
http//visualparadox.com/wallpapers/binary1600.ht
m
7
Data Gatherer

• Data variables
• Metadata
• Audit trails
• Analytical methods
• Publications

8
Relevance Algorithms

• Data to Data
• User to Data
• Methods to Data

9
Sociometric Network Analysis

• Community detection
• Clustering

10
(No Transcript)
11
(No Transcript)
12
Benefits
13
Assisted Metadata
Benefits

• Machine learning algorithm

14
Human Input

• Interaction with archivist and researchers

15
Standardized APIs

• Designed to work with repositories

16
Guided Metadata Creation

• Assist with keyword selection
• Suggest ontologies

17
Linking Data

• Across data types
• Across disciplines

18
Recommending Data

• Collaborative filtering
• Multi-disciplinary team suggestions

Slide acknowledgment Wikimedia Commons
http//psychosocial.actalliance.org/default.aspx?d
i65446ptid66401
19
Thank You
Jonathan_Crabtree_at_unc.edu
http//www.odum.unc.edu
http//databridge.web.unc.edu