Understanding the Rhythms of Relationships in Email Archives

1
Understanding the Rhythms of Relationships in
Email Archives

• Research by Adam Perer, Ben Shneiderman, Douglas
  Oard
• Institute for Advanced Computer Studies
• Presented by Susan E. Davis, College of
  Information Studies
• University of Maryland

2
Motivation
Email archives are important artifacts for
understanding the individuals and communities
they represent
3
Typical Email Browser
4
University of Maryland Projects

• Enron email project
• Joint Institute for Knowledge Discovery
• http//zaphod.mindlab.umd.edu16080/JIKD/index.sht
  ml
• http//jikd-email.umiacs.umd.edu/index.html
• Digital Archive of the Dot Com Era
• Robert H. Smith School of Business/NDIIPP,
  Library of Congress
• http//www.dotcomarchive.org

5
Previous Work
6
Shneiderman Archive

• 44,971 saved messages
• 4,051 relationships

7
Relationships
8
Rhythms of Relationships
9
Distribution of Relationships
Number of Messages
Relationships
10
Querying by Rhythm
11
Most Active Professional Relationships

• TYPE NUMBER AVE YRS AVE
• UMD-Close colleagues 11 9.2 209.7
• UMD-Superiors staff 11 9.6 123.0
• UMD-Students 9 9.0 183.8
• Colleagues at other Univ. 17 11.3 152.4
• Conference partners 10 8.3 172.7
• Corporate partners 9 9.1 137.6
• USACM Public Policy 4 5.5 252.3
• Book editorial workers 3 8.7 183
• Government partners 2 9.5 171.5

12
Growth Rates for Messages and Relationships
13
Over 4,000 Relationships Superimposed
14
k-means Clustering
Spotfire http//www.spotfire.com/
15
Zoomed-in view of the dendrograms 1988 subtree
16
Aggregating Related Rhythms

• Organizations (user_at_ibm.com)
• Organization Type (user_at_umd.edu)
• Country Codes (user_at_cam.ac.uk)
• Collaborations

17
Aggregate Rhythms
18
Kenneth Lays (Enron Chairman/CEO) Rhythms of
Relationships
19
Future Work

• Clustering independent of time
• Granularities of months, weeks, days
• Integrating external contextual clues
• Calendars, Conference Dates, Lectures

20
Lessons Learned

• Visualizations are useful for analysis and
  sensemaking tasks
• Enables explorers to
• Analyze large and long archives
• Find patterns of interest
• Identify aggregate trends
• Provides needed context

http//www.cs.umd.edu/hcil/emailviz/