Visualizing Network Data

1
Visualizing Network Data

• Richard A. Becker
• Stephen G.Eick
• Allan R.Wilks

2
Network Data

• Understanding network data is crucial
• Dataset
• Internet data, telephone data
• www, e-mail communication
• Challenge coping with data volumes

3
SeeNet

• Goal visualizing the network data, not just the
  structure of the network itself
• Reduce the amount of data
• Aggregation, Averaging, Thresholding, Exception
  reporting
• Techniques Static displays, Interactive
  controls, Animation

4
Network Displays

• Linkmaps
• Map clutter problem
• Nodemaps
• Use aggregation
• Omit detailed information
• Matrix Display
• In/out nodes are assigned in rows/cols
• Omit geographical information

5
Linkmaps vs. Nodemaps
6
Parameter Focusing

• Select the parameter values controlling the
  characteristics of the display
• Parameter Values
• Statistic, Levels, Geography/Topology, Time,
  Aggregation, Size, Color..
• Issues
• Large space of possible values
• Most combinations are not understandable
• Displays are sensitive to particular values

7
Interactive Control

• Identification
• Linkmap/Nodemaps
• Sliders for line length, thickness, animation
  speed, color, time, symbol size
• Matrix Display
• Permutation of the rows/cols(drag-and-drop)
• Zooming and Birds-Eye
• Animation/Conditioning/Sound

8
Interactive Control(link length)
9
Further Examples

• CICNet
• Packet-switched data network
• Schematic, not geographic
• E-mail Communications
• Nodes(users) and Links(emails)
• World Internet
• Display statistics from the internet

10
Further Examples(2)

• CICNet

• E-mail Communication

11
My Favorite Sentences

• To solve the display clutter problem, we invented
  a suite of parametric techniques embodied in a
  dynamic graphics software system called SeeNet
  that enable a user to focus the display and
  thereby reveal patterns in the network data

12
Contribution

• Focus on both a network and data on that network
• Try to solve the display clutter problem
• Dynamic parameter control

13
Reference

• Fairchild, Poltrock, Furnas15
• the SemNet system
• Sarkar, Brown16
• Visualizing the structure of sparse networks
• Paulish17
• Edge concentration, gradual focusing
• Ahlberg, Shneiderman34
• A nearly instantaneous response is critical

14
Critique

• Strength
• Three different display methods
• Lots of parameters that users can choose
• Easy to manipulate parameters
• Can produce good visualization for various
  network datasets
• Weakness
• Is it easy to find the best combination of
  parameters and display method?

15
What has happened to this topic?

• Constructing Interactive Network Visual
  Interfaces (1998)  
• Cited by
• CyberNet A framework for managing networks using
  3D metaphoric worlds  
• Real-Time Geographic Visualization of World Wide
  Web Traffic (1996)  
• CAIDA visualization tools
• http//www.caida.org/tools/visualization/