Docstoc

Usability _ Research Frontiers

Document Sample
Usability _ Research Frontiers Powered By Docstoc
					                                                                                                    Toronto




                                                                          The New Yorker, September 2001


Christopher Collins • CSC 318                                                      15 April 2006 • 1




                                                                                                    Toronto


  Information Visualization & Visual Design:
  Usability & Research Frontiers
                                                             CSC 318, Fall 2005

                                Christopher Collins (ccollins@cs.utoronto.ca)

                                Department of Computer Science

                                University of Toronto

                                14 March 2006
                Acknowledgement: Where noted material is based presentations by Ilona Posner & Ron Baecker.




                                                                                                              1
       History of Information Visualization                                              Toronto


                                Usability and Usefulness

                                     Visual Variables

                            Choosing a Representation

                                  Graphical Integrity

                                       Design Pitfalls

                                  Research Frontiers

Christopher Collins • CSC 318                                            15 April 2006 • 3




                                                                                         Toronto




                                                      Advertisem   ent
                                  Royal Baking Powder




Christopher Collins • CSC 318                                            15 April 2006 • 4




                                                                                                   2
  Information Visualization
                                                                                    Toronto

   • Compact graphical presentation and user interface for
            manipulating large numbers of items (102 - 106),
            possibly extracted from far larger datasets.

   • Enables users to make
            discoveries, decisions, or explanations

   • about
            patterns (trend, cluster, gap, outlier...),
            groups of items, or
            individual items.
                                                           (Ben Shneiderman, UMD)


Christopher Collins • CSC 318                                       15 April 2006 • 5




  5 Advantages of Visualization
   • Comprehension                                                                  Toronto


            enhances capacity and efficiency of data comprehension

   • Perception
            reveals unexpected properties of the data

   • Quality Control
            makes problems in data or data collection apparent

   • Focus + Context
            facilitates understanding of small scale in context of full picture
            of data

   • Interpretation
            supports hypothesis formation, leading to further investigation
                                                                            (Ware, 2000)
Christopher Collins • CSC 318                                       15 April 2006 • 6




                                                                                              3
  Applications of Visualization
   • Banking and finance                               Toronto



   • Industry and manufacturing

   • Resources and exploration

   • Science and medicine

   • Education

   • Software development

   • Military (Intelligence)                  (Ilona Posner)

Christopher Collins • CSC 318          15 April 2006 • 7




  History
• Long history of visual                               Toronto

  communication:
        Illustration
        Gesture
        Sign languages
        Emoticons
        Pictographic writing systems
        Statistical graphics

• Computers allow us to
  automate creation of visual
  representations of data


Christopher Collins • CSC 318          15 April 2006 • 8




                                                                 4
  Example: Movement of the French Army
                                                                                 Toronto




                                                     Charles Joseph Minard (1861)
     temperature, date               number location, direction
Christopher Collins • CSC 318                                    15 April 2006 • 9




  London Cholera Deaths
                                                                                 Toronto




 •   John Snow (1855) On the Mode of Communication of Cholera.
Christopher Collins • CSC 318                                    15 April 2006 • 10




                                                                                           5
  More Relevant Today …
   • Computer science involves the study of data management and                       Toronto
     data processing.
   • Manage and process data by computers because:
             data sets are large;
             data sets change over time;
             data sets embody complex interrelationships.
   • But how do we comprehend the data?
            How do we know what is there and what is missing?
            How do we know if there are errors in the data?
            How do we know what relationships exist in the data?
            How do we know what “the data means”? - We usually don’t!
   • We can't see/visualize/perceive/understand:
     Good data display and visualization techniques can help!
                                                                   (Ilona Posner, 2005)

Christopher Collins • CSC 318                                          15 April 2006 • 11




  Example: Authorship and Style
                                                                                      Toronto

   • Letter tuple frequency (3 AB, 5 TH, …) mapped to 2D position


      The Federalist Papers




        Hamilton                    Madison                   Disputed
                                                                (Kjell et al., 1992; 1994)


Christopher Collins • CSC 318                                          15 April 2006 • 12




                                                                                                6
                      History of Information Visualization
                                                                                                 Toronto



                    Usability and Usefulness
                                        Visual Variables

                            Choosing a Representation

                                     Graphical Integrity

                                          Design Pitfalls

                                     Research Frontiers
Christopher Collins • CSC 318                                                     15 April 2006 • 13




  Usability of Visualization
                                                                                                 Toronto
   • Data Design
            Select the right data dimensions
            Pitfall: Display irrelevant data relationships

   • Visual Design
            Consider perceptual capabilities
            Pitfall: Difficult to interpret, lead users to misinterpretation

   • Interface Design
            Mode of interaction with visual and data appropriate
            Pitfall: Poor interaction negates benefits from data and visual design

   • Understanding Target Users
            Visualization design accounts for stakeholder needs and characteristics
            Pitfall: Mistaken assumptions of cultural norms or user abilities leads to misinterpretation



Christopher Collins • CSC 318                                                     15 April 2006 • 14




                                                                                                           7
  Data Design
                                                                       Toronto

   • Usually guided by identification of a problem data set
     for which interpretation is currently difficult.

   • Visualization designer must work with target audience.

   • Visualization is one of many types of ways to display
     relationships: like statistics, poor visualization design can
     lead to apparent relationships between just about
     anything… if possible, confirm data dimensions are
     indeed correlated before beginning.




Christopher Collins • CSC 318                           15 April 2006 • 15




  Visual Design
                                                                       Toronto


• Theory of how to choose a visualization in its infancy.

• One theory advises a process of matching visual variables
  with data dimensions.

• Visual variable’s usable dimensionality related to preattentive
  characteristics, compositionality, and clever use of
  dimensional matching to reduce cognitive load.

• Design with Graphical Integrity in mind




Christopher Collins • CSC 318                           15 April 2006 • 16




                                                                                 8
  Graphical Excellence
   • Principles of graphical excellence                                       Toronto



            Complex ideas presented with clarity, precision,
            honesty, and efficiency.
            Gives viewer the greatest number of ideas in the
            shortest time, with the least ink in the smallest space.

   • Graphical excellence often found in simplicity
     of design and complexity of data.
            Recall Minard’s map; multivariate data, simple
            design
   (Edward Tufte, Visual Display of Quantitative Information, 1983, 2001)


Christopher Collins • CSC 318                                  15 April 2006 • 17




  Emergeny Center: Telling a Story
                                                                              Toronto




                                                              Visual Correlation
                                                              for Situational
                                                              Awareness

                                                              (Livnat et al., 2005)


Christopher Collins • CSC 318                                  15 April 2006 • 18




                                                                                        9
  High Dimensionality Doesn’t Guarantee Excellence
                                                                                                                       Toronto




                                                                                         (Chlan and Rheingans, InvoVis 2005)

Christopher Collins • CSC 318                                                                           15 April 2006 • 19




  Reduce Learning Curve: Small Multiples
                                                                                                                       Toronto




                                                       scuttle
                       slop



                                          swish                  romp
                 thunder
                                                                            scramble

   squeal

                                                   clatter        journey
                                      lurch
                                                                                limp

            screech

                                              gallop                        hobble
                                                             stumble
                inch

                           stagger
                                                                                trudge
                                                       stroll
                                shuffle




       (Nina Thiessen, 2004)

Christopher Collins • CSC 318                                                                           15 April 2006 • 20




                                                                                                                                 10
                      History of Information Visualization
                                                                          Toronto

                                Usability and Usefulness


                                Visual Variables
                            Choosing a Representation

                                  Graphical Integrity

                                     Design Pitfalls

                                  Research Frontiers
Christopher Collins • CSC 318                              15 April 2006 • 21




  Perception Review
                                                                          Toronto

   • (Mis)interpretation depends on human
     perception.




Christopher Collins • CSC 318                              15 April 2006 • 22




                                                                                    11
  Visual Variables
                                                            Toronto




                                                  (Bertin, 1983)

Christopher Collins • CSC 318                15 April 2006 • 23




  Visual Variables
                                                            Toronto




                                (Bertin, 1983; Carpendale, 2003)

Christopher Collins • CSC 318                15 April 2006 • 24




                                                                      12
  Visual Variables
   • Others that have not been studied:                                        Toronto


            overlay (transparency)
            time (e.g. chronological order of appearance)
            transformation (morphing, animation)

   • Not all are equally powerful – position, colour better than
     size, shape (which can be confounded by illusions).
            Map most important feature to position!

   • How many can be used in conjunction? Do the
     characteristics change in combination?
            Work of Colin Ware; more study is needed


Christopher Collins • CSC 318                                   15 April 2006 • 25




  Interpretation Bias: Power of Colour
                                                                               Toronto




                                                       (Bernice Rogowitz, 1995)
Christopher Collins • CSC 318                                   15 April 2006 • 26




                                                                                         13
                       History of Information Visualization
                                                                                           Toronto

                                Usability and Usefulness

                                      Visual Variables


                Choosing a Representation
                                  Graphical Integrity

                                       Design Pitfalls

                                  Research Frontiers
Christopher Collins • CSC 318                                               15 April 2006 • 27




  Review: Representations
• A representation is
                                                                                           Toronto
        a formal system or mapping by which the information can be specified
           (D. Marr)
         a sign system in that it stands for something other than its self.
• for example: the number thirty-four
           decimal: 34,
           binary:  100010,
           roman: XXXIV

• different representations reveal different aspects of the information
           decimal: counting & information about powers of 10,
           binary:  counting & information about powers of 2,
           roman: counting

• presentation is not representation
           how the representation is placed or organized on the screen

             34,                 34                          (Edward Tufte (2001), Ilona Posner)

Christopher Collins • CSC 318                                               15 April 2006 • 28




                                                                                                     14
  Representations
                                                                       Toronto


• Solving a problem simply means representing it so
  as to make the solution transparent (Simon, 1981)
• Good representations
        allow people to find relevant information
        information may be present but hard to find
        allow people to compute desired conclusions
        computations may be difficult or “for free” depending on
        representations (long division using decimal vs. roman
        numerals)
                                                        (Edward Tufte (2001)

Christopher Collins • CSC 318                           15 April 2006 • 29




  Dimensional Matching
                                                                       Toronto

   • Choice of visual variable to map to data
     dimension based on:
            size & properties of variable
            strength of variable (position > colour > size > …)
            match of data dimension to visual dimension




Christopher Collins • CSC 318                           15 April 2006 • 30




                                                                                 15
  Representation in Visualization Design:
  Dimensional Matching
                                                                       Toronto




   0                             20

           Adults               Calves
                                         (Buffalo Image: Edward Tufte (2001))
Christopher Collins • CSC 318                           15 April 2006 • 31




  Add additional data dimensions…
                                                                       Toronto

   • Number of adults and calves +

   • Horizontal buffalo distribution

   • Sky colour in photo




Christopher Collins • CSC 318                           15 April 2006 • 32




                                                                                 16
  Dimensional Matching
                                                      Right                  Toronto




                                                      Left


   0                             20

           Adults               Calves       Sky
Christopher Collins • CSC 318                                 15 April 2006 • 33




  Dimensional Matching
                                                                             Toronto
  • Matched version more memorable & easier to read

  • Both are "useful", only one "usable"

  • How many total buffalo? Overlapping?                 Right
  • Apophenia? Buffalo prefer lighter sky!




                                                         Left


Christopher Collins • CSC 318                                 15 April 2006 • 34




                                                                                       17
  Avoiding Misinterpretation of Dimensions
                                                                    Toronto


                                  • If dimensions don’t match
                                    expectations, provide a
                                    legend!




                  3D Mars Globe

                                   elevation
   (Zoran Constantinescu)
Christopher Collins • CSC 318                       15 April 2006 • 35




    Choosing a Representation: Summarization

                                                                    Toronto




                                                    (Leuski et al., 2003)
Christopher Collins • CSC 318                       15 April 2006 • 36




                                                                              18
                      History of Information Visualization
                                                                          Toronto

                                Usability and Usefulness

                                    Visual Variables

                            Choosing a Representation

                                Graphical Integrity
                                     Design Pitfalls

                                  Research Frontiers

Christopher Collins • CSC 318                              15 April 2006 • 37




  Edward Tufte: “Graphical Integrity”
                                                                          Toronto
   • Minimize “ink” while maximizing data density

   • Avoid “chart junk” – embellishments

   • Prefer multi-functioning graphics, parallelism
     (small multiples)

   • Do not map data to higher dimension
            Example: Don’t map length to volume of a sphere




Christopher Collins • CSC 318                              15 April 2006 • 38




                                                                                    19
  Integrity
                                                                            Toronto




990? 4,000?2,000?
                                       ?


                                                      (Macleans, 2005)

Christopher Collins • CSC 318                                15 April 2006 • 39




  Lie Factor 14.8
                                                                            Toronto




   • Comparing line length or area? Perspective distorts difference!
                                                               (Tufte, 2001)

Christopher Collins • CSC 318                                15 April 2006 • 40




                                                                                      20
                                                Toronto




                                   (Tufte, 2001)

Christopher Collins • CSC 318    15 April 2006 • 41




  Integrity without Excellence
                                                Toronto




                                   (Tufte, 2001)

Christopher Collins • CSC 318    15 April 2006 • 42




                                                          21
  Importance of Context
                                               Toronto




                                  (Tufte, 2001)

Christopher Collins • CSC 318   15 April 2006 • 43




  Importance of Context
                                               Toronto




                                  (Tufte, 2001)

Christopher Collins • CSC 318   15 April 2006 • 44




                                                         22
  Importance of Context
                                               Toronto




                                  (Tufte, 2001)

Christopher Collins • CSC 318   15 April 2006 • 45




  Importance of Context
                                               Toronto




                                  (Tufte, 2001)

Christopher Collins • CSC 318   15 April 2006 • 46




                                                         23
  Chart Junk & Fontitis
                                                                                      Toronto
                 10

    10

                                      8

     8
                                                  Dear Sir,

     6
                                                  This is a really exciting
                                  5
                                                    opportunity!
     4

                         2
                                                  Take advantage of it
     2
                                                    NOW!

     0


                                                                           (Ilona Posner)

Christopher Collins • CSC 318                                          15 April 2006 • 47




  Removing Deception: Baseline
                                                          Maintenance cost / year     Toronto
         66

         65

         64

         63

         62

         61

         60

         59

         58
              Ford           GM       Pontiac   Toyota


                                                                           (Ilona Posner)

Christopher Collins • CSC 318                                          15 April 2006 • 48




                                                                                                24
  Removing Deception: Scale
                                                                   Ford             Toronto
                                                                   GM
                                                                   Pontiac
                                                                   Toyota




                                                                          (Ilona Posner)

Christopher Collins • CSC 318                                        15 April 2006 • 49




  Removing Deception: 3D Perception
                                                        Maintenance cost / year     Toronto
       70
       65
       60
       55
       50
       45
       40
       35
       30
       25
       20
       15
       10
        5
        0
               Ford             GM   Pontiac   Toyota

                                                                          (Ilona Posner)

Christopher Collins • CSC 318                                        15 April 2006 • 50




                                                                                              25
  Removing Deception: Simplify
                                                                                                                                                   Toronto
      70
                                                                                                Maintenance cost / year
      65
      60
      55
      50
      45
      40
      35
      30
      25
      20
      15
      10
          5
          0
                     Ford               GM           Pontiac                    Toyota

                                                                                                                                    (Ilona Posner)

Christopher Collins • CSC 318                                                                                               15 April 2006 • 51




  Removing Deception: Simplify
                                                          Maintenance cost / year                                                     Ford
     66
                                                                                                                                      GM           Toronto
                                                                                                                                      Pontiac
     65
                                                                                                                                      Toyota

     64

     63

     62

     61

     60

     59



                                        Ford                         GM                         Pontiac   Toyota
     58
              Ford     GM     Pontiac      Toyota



              Maintenan                                  60                          61                59        66
                                                    Maintenance cost / year         70
      70                                                                                                                 Maintenance cost / year
                                                                                    65
      65
                                                                                    60
      60
                                                                                    55
      55
                                                                                    50
      50
                                                                                    45
      45
                                                                                    40
      40
      35                                                                            35
      30                                                                            30
      25                                                                            25
      20                                                                            20
      15                                                                            15
      10                                                                            10
       5                                                                             5
          0                                                                          0
              Ford     GM   Pontiac     Toyota                                           Ford    GM   Pontiac   Toyota


                                                                                                                         (Credit Ilona Posner)

Christopher Collins • CSC 318                                                                                               15 April 2006 • 52




                                                                                                                                                             26
                      History of Information Visualization
                                                                          Toronto

                                Usability and Usefulness

                                    Visual Variables

                            Choosing a Representation

                                  Graphical Integrity


                                Design Pitfalls
                                  Research Frontiers
Christopher Collins • CSC 318                              15 April 2006 • 53




  Design Pitfalls
                                                                          Toronto

   • Visual Design
            Avoid “interface junk” – embellishments (e.g. fancy
            icons) which detract from the data visualization.

   • Interface Design
            Consider 3D carefully, can actually reduce
            performance (obscuring, difficulty navigating space
            with a mouse).
            Choice of physical interface: may be better on a
            tabletop display than a wall, or better on a wall-sized
            display than a cell phone.

Christopher Collins • CSC 318                              15 April 2006 • 54




                                                                                    27
                                                                          Toronto




                 Kartoo.com: visual search with interface junk 2006 • 55
Christopher Collins • CSC 318                              15 April




                                                                          Toronto




Christopher Collins • CSC 318                                15 April 2006 • 1995)
                                                           (Marti Hearst,56




                                                                                     28
                                                             Toronto




Christopher Collins • CSC 318                 15 April al., 57
                                           (Rohrer et2006 • 1999)




  2D Gist Icons
                                                             Toronto




                                (DeCamp et al, 2005)


Christopher Collins • CSC 318                 15 April 2006 • 58




                                                                       29
  Consider Target Audience
                                                                                        Toronto




                                            (Geoff Coulson, Meteorological Service of Canada)

Christopher Collins • CSC 318                                            15 April 2006 • 59




                      History of Information Visualization
                                                                                        Toronto

                                Usability and Usefulness

                                    Visual Variables

                            Choosing a Representation

                                  Graphical Integrity

                                     Design Pitfalls


                           Research Frontiers
Christopher Collins • CSC 318                                            15 April 2006 • 60




                                                                                                  30
  Visualization “Mantra”
                                                                                             Toronto




       Overview first, zoom and filter, then details on demand.
                                (Shneiderman, Designing the User Interface 3rd Ed. 1997 p523)




                   Gallery: www.visualcomplexity.com




Christopher Collins • CSC 318                                                 15 April 2006 • 61




                                                                                             Toronto




Christopher Collins • CSC 318                                                 15 April 2006 • 62




                                                                                                       31
                                               Toronto




Christopher Collins • CSC 318   15 April 2006 • 63




                                               Toronto




Christopher Collins • CSC 318   15 April 2006 • 64




                                                         32
  Elastic Presentation
                                                   Toronto




                                (Carpendale, 1999)

Christopher Collins • CSC 318       15 April 2006 • 65




  FlowMap Layout
                                                   Toronto




       (Phan et al., 2005)


Christopher Collins • CSC 318       15 April 2006 • 66




                                                             33
  Vizter: Social Network Visualization
                                                                        Toronto




 (Heer and Boyd, 2005)
Christopher Collins • CSC 318                            15 April 2006 • 67




  Information Visualization in Everyday Use
                                                                        Toronto




   • DateLens (Bederson & Clamage, 2003, CHI 2003 DVD)

Christopher Collins • CSC 318                            15 April 2006 • 68




                                                                                  34

				
DOCUMENT INFO