MAT 259 Visualizing Information

1
MAT 259 Visualizing Information

• Winter 2006, e-studio, Art 2220
• Tues 1000-1200, Lecture
• Thurs 1000-1200, Lab
• George Legrady, legrady_at_arts.ucsb.edu
• TA Angus Forbes, angus.forbes_at_gmail.com
• Course Web Site http//www.mat.ucsb.edu/g.legra
  dy/ (click on courses, click on MAT 259)

2
Course Goals

• An introduction to information visualization
• An overview of varied methodologies
• Comparison between uses in diverse disciplines
• Introduction to self-organizing algorithms
• Project driven course with focus on theory and
  practice
• Working with cultural data,
• Exploration of methodologies,
• Visualization output to reflect aesthetic
  consideration

3
Workload

• Attendance at weekly lectures
• Active participation
• Online reports on readings
• Attendance reports on visiting lectures
• Completion of warm-up and final projects

4
Visualization Cross-Disciplinary Fertilization

• Domain visualization, an emerging field
• Multi-disciplinary Difficult to get the overview
  of the field
• Researchers bring their own disciplines
  perspective
• Examination of other disciplines export and
  import of methods, ideas, models, or empirical
  results
• Creative imagination required to foresee how
  outside info fits the problem at hand

5
Discipline Driven Methodologies

• Each discipline has a particular implementation
  goal
• LSIS citation indexing, bibliographic indexing,
  etc.
• Scientific Visualization Map physical phenomena
  in 2D, or 3D
• Information Visualization Analyzing and
  transforming nonspatial data into visual form
• Geographic Information Systems (GIS)
  Cartographic framework, a familiar way to map
  data
• Art Aesthetics, complexity, culturally
  meaningful results

6
Goal Driven Methodologies

• Information Visualization visually map abstract,
  nonspatial info
• Information retrieval research in vast data sets
• Depicting the overall semantic structure of a set
  of documents
• Identifying patterns through visualization (DNA)

7
User Meta Model

• Data Extraction
• Definition of Units of Analysis
• Selection of Measures
• Calculation of similarity between units
• Ordination assignments of coordinates to each
  unit
• Analysis and Interpretation of output
  visualization

8
Classification Methods

• Alphabetical arbitrary learned system
• Numeric arbitrary learned system
• Scalar (hotel star system) implies value scale
• Sequential (time) based on units
• Spatial sense of place
• Categories similar things grouped together
• Associative (If a to b, then c to d)
• Metaphoric A way to establish context
• Random Creates complexity (game beginnings)

9
Visualization Process

• Multivariate data to be presented in 2D in print
  or computer screen
• by applying mathematical dimensionality
  algorithms to map the data
• Clustering techniques to group similar data
• Spatial proximity matrix similar data/close,
  difference/distance
• Large amounts of data presented in limited space
  
• Panning, zooming, filtering to access data

10
What is Visualization?

• Design of the visual appearance of data objects
  and their relationships
• Ability to comprehend large amounts of data
• Reduction in search time through visualization
• Provides a better understanding of complex data
  sets
• Reveal relationships and properties through
  visual perception
• Multiple simultaneous perspectives
• Effective communication

11
Formal Aesthetic Functions

• Visualization Design years of expertise and
  diverse skills
• Visual communication a language system (function
  of form, colors, etc)
• Complex data relationship benefit from
  storytelling
• Narrative methods enhance communication

12
Interaction Design

• Search and browse through data
• Zoom, filtering, panning, etc.
• 1) Overview, 2) Zoom-in (filter), 3)
  Details-on-demand
• Browsing explores both the organization or
  structure of the information space, and its
  content (Chen, 1998)
• Information architects design layered info spaces
  based on classification systems
• 3 Navigational Paradigms 1) spatial, 2)
  semantic, 3) social (using behavior of
  like-minded people) (Dourish)

13
Visualization Outcomes

• Effective exploitation of perceptual principles
• Helps communication with non-specialists
• Discover hidden (semantic) patterns, structures
• Contribute to knowledge development in all
  disciplines

14
References (Selected)

• Visualizing Knowledge Domains, Borner, Chen,
  Boyak
• Journal of Information Visualization
• Kohonen Self-Organizing Algorithm
• Visual Complexity
• Information Aesthetics
• Edward Tufte