Visualization Encoding by yco10525

VIEWS: 5 PAGES: 31

									Visualization Encoding
Introduction
 Information visualization starts from data.
 There are many forms that the data could
  take, text, spreadsheets, relational DB tuples,
  etc.
 There are many patterns that the data could
  follow, clustering, outlier, correlation, etc.
 Encoding:
              Application
       Data    Domain         Graphic
                            Presentation
Fundamental Tasks
 Information presentation.
    Maps, Photographs, Movies, …
 Information extraction.
    Interactive graphical interface
Information Presentation
Data Mining Example: Clustering
Information Extraction
Data Mining Example: Clustering
Data Types
 1-D, 2-D, 3-D, temporal, multi-dimensional,
  tree and network data.
 Data types characterize the information
  objects in the task domain.
Basic Visualization Tasks
 Overview of a collection of data.
 Zoom in/on objects of interest.
 Filter out uninterested items.
 Details-on-demand: view details.
 Relate: View relationship.
 History: Undo, Redo, Refinement.
 Extract a subset of the data.
1-D Data and Task Encoding
 Linear data: textual document, source code,
  etc.
 User problems: count, find, replace, …
 Encoding: fonts, color, size, layout,
  scrolling, selection capabilities, …
 Product example: text editor, browser, …
2-D Data and Task Encoding
 Planar or map data: geographical maps,
  floor plans, newspaper layouts, …
 User problems: find adjacent items, search
  containment, find paths, filtering, details-
  on-demand, …
 Encoding: size, color, layout, arrangement,
  multiple layers, …
 Product example: CAD
3-D Data and Task Encoding
 Real-world objects: building, human body
 User problems: adjacency in 3-D,
  inside/outside relationship, position,
  orientation, occlusion
 Encoding: overviews, landmarks,
  transparency, color, perspective, stereo
  display
 Product example: CAD
Temporal Data and Task
Encoding
 Time series data: medical records, project
  management, historical presentation
 User problems: finding all events before,
  after or during some time period or moment.
 Encoding: time lines
Multi-dimensional Data and Task
Encoding
 Relational and statistical databases tuples.
 User problem: finding patterns, clusters,
  correlations, gaps, outliers.
 Challenge:
  – Simultaneously display many dimensions of
    large subsets of data.
  – Create displays that best encode the data pattern
    for a particular task.
  – Rapidly select a subset of tuples or dimensions.
An Encoding Example
Dimensionality Encoding
 Multi-dimensional databases are structured
  as n-dimensional data cube.
 The dimensions of the data can be explicitly
  encoded in the structure of tables.
Data Set Encoding
 The data sources are encoded as layers.
 The different result sets are encoded as
  different panes in different layers.
User Interest Encoding
 Providing enough tools and allowing user to
  specify his interest.
 The table configuration encodes the user
  interest.
 Table configurations are defined in form of
  algebra
  – Concatenation
  – Cross product
  – Nest (Division)
 For ordinal fields, algebra operand symbols
  take all domain values.
  – A = domain (A) = {a1, a2, …, an}
  – Example: Month = {Jan, Feb, …, Dec}
 For quantitative fields, algebra operand
  symbols take the field names as values.
  – P = {P}
  – Example: Profit = {Profit}
 Ordinal fields partition the table into rows
  and columns; quantitative fields are
  spatially encoded as axes within the panes.
 Concatenation Example:
  – Quarter = {Qtr1, Qtr2, Qtr2, Qtr4}
  – Product = {Coffee, Espresso, Herbal, Tea}
  – Profit = {Profit}, Sales = {Sales}




  Ordinal Field        Group By




  Quantitative Field       Sorted By
 Cross Product Example:
  – Ordinal x Ordinal




   Ordinal x Quantitative
 Nest (Division) Example:




   Quantitative field does not make sense for
    divisions
Product x SumOfSales   Quarter x SumOfProfit
Types of Graphics inside Panes
 Types of Panes:
  – Ordinal – Ordinal
  – Ordinal – Quantitative
  – Quantitative - Quantitative
Visual Encoding
 Shape
 Size
 Orientation
 Color
Tree Type Data and Task
Encoding
 Exponential data: hierarchies, tree
  structures.
 User problems: find the structural properties
   – Height of the tree
   – Number of children
   – Find nodes with same attributes
 Encoding:
  – Outline style of indented labels
    Node-link diagrams: allowing the encoding of
     linkage between entities.
    Treemap: child rectangles inside parent
     rectangles
 Product example: windows explorer, internet
  traffic, hyperbolic browser
Network Data and Task
Encoding
 Graph data: multiple paths, cycles, lattices
 User problems:
  – Shortest path
  – Topology problems
 Encoding: imperfect
  – Node-link diagram
  – Matrix
General Encoding Principles
 Expressiveness:
  – Encode all the facts in the result set.
  – Encode only the facts in the result set.
 Effectiveness:
  – Depends on the capability of the perceiver.
  – Encode the more important information more
    effectively.
– Perceptual accuracy ranks
Conclusion
 Visualization helps
  – Information presentation
  – Information extraction
 Good visual encoding should match the
  target data and user problems.
 Studying the successful/unsuccessful visual
  encoding designs and techniques helps us to
  design and develop new encoding
  approaches.

								
To top