Visualization

Two branches, not always clearly separated:

Information visualization ("cognition amplification")
Scientific visualization (engineering, medicine, genetics)

Everyday application: disk analyzers such as SequoiaView.

Hot topics:

KDD: Knowledge discovery in databases
DM: Data Mining
VDM: Visual Data Mining

Two uses of visualization:

Exploratory (finding patterns or rules interactively)
Presentational (communicating them as models, often interactively)

Survey paper: Oliveira, Levkowitz, "From Visual Data Exploration to Visual Data Mining: A Survey", IEEE Transactions on Visualization and Computer Graphics 9(3), pp. 378-394 (2003). Freely available from SUUB from PCs inside the network of Bremen's universities. (Note the submission date and the printing date; such a time span is typical.)

Types of data:

Tables, databases of few or many dimensions (independent variables) or variates (dependent variables)
Networks, graphs, hierarchies
Streamlines, surfaces, geometric propertiees
Algorithms, processes
Text, multimedia, CAD
...

Classical visualization techniques: Keim, "Visual Techniques for Exploring Databases"

Visualization can be modeled as a pipelined process: Chi, Riedl, "An Operator Interaction Framework for Visualization Systems"

Visualization in Networking

Example: Ball, Fink, North, "Home-Centric Visualization of Network Traffic for Security Administration". For lots of other examples see CAIDA.

Visualization in networking is mostly concerned with drawing graphs nicely, that is: in a visually effective and efficient manner ("graph layout").

We may not always use straight lines: Hyperbolic trees use non-Euclidean geometry to zoom in on a region of interest. There are many more ways of nonlinear magnification.

Typical way of layout: Define a physical "energy" to create repelling and attracting spring forces that push the graph nodes to a nice configuration.