Subjective Quantification of Perceptual Interactions among some 2D Scientific Visualization Methods

1
Subjective Quantification ofPerceptual
Interactionsamong some 2DScientific
Visualization Methods
Daniel Acevedo David Laidlaw
Visualization Research Lab Brown
University http//vis.cs.brown.edu
2
Icon Size
Icon Spacing
Icon Brightness
3
Our Goal

• Model the effectiveness of multi-valued,
  multi-layered, icon-based
• 2D scientific visualization methods
• Quantify the perceptual interactions among
  visual dimensions
• Quantify the expressive power of the
  individual visual dimensions

4
Paper Summary
Quantifying filtering interference among 3
visual dimensions when visualizing a
single-valued continuous scalar dataset in 2D
for exploratory visualization using icon-based
visualization methods.
5
Roadmap of the talk

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

6

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

• Info/Sci. Visualization
• Bertin, Cleveland, et al., Mackinlay, Card,
• MacEachren,
• Perception
• Callaghan, Landy, Bergen, Carswell and
  Wickens, Ware, Healey,
• Bair, House, Interrante,
• Visual Design and Art
• Wallschlaeger et al., Tufte,

7

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

8

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Exploratory Visualization

• Prompt visual thinking
• Discovery and hypothesis formation
• Overall understanding
• Insight
• Design factors
• Data characteristics
• Composition

Spatial resolution Data levels Context and
saliency Data relations Visual clutter Balance
9

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

The Task Exploratory Visualization
10

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Summary

• Want to
• Quantify expressive power of visual dimensions
• Have
• Design factors to qualify our exploratory goal
• Need
• More indirect experimental techniques

11

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Logistics

• 3 Visual Dimensions
• Icon Brightness
• Icon Size
• Icon Spacing

Mapped Icon size (0,1) Non
Mapped Icon brightness (1.00) Icon spacing
(0.66)
Filtering Interference Brightness x Size x
Spacing Size x Spacing Spacing x Size
12

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Logistics

• 3 Tasks
• Data Resolution Identification
• How many different levels of data can a visual
  dimension represent?
• Spatial Feature Resolution Identification
• What is the size of the smallest
  spatial feature a visual dimension can
  represent?
• Visual Linearity Perception
• How perceptually linear is a
• visual dimension?

13

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Logistics

• Computer-based. 900x900.
• 6 Subjects. Within-subjects design.
• 9 training sections, 9 trial sections.
• No time limit, with breaks.

14

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
How many different levels of datacan a visual
dimension represent?
D zero-crossing
(a, l)
Da

• a 0.2, 0.6
• l 0.625, 1.25, 2.5, 5 ( image width)

15

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Icon brightness
16

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Icon brightness
17

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Icon size
18

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Icon size
19

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Icon spacing
20

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Icon spacing
21

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
What is the size of the smallest spatialfeature
a visual dimension can represent?
Dl
22

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon brightness
23

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon brightness
24

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon size
25

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon size
26

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon spacing
27

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon spacing
28

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
How perceptually linearis a visual dimension?
29

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon brightness
30

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon brightness
31

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon size
32

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon size
33

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon spacing
34

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon spacing
35

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Summary

• 3 Visual Dimensions
• Icon Brightness
• Icon Size
• Icon Spacing
• 3 Tasks
• Data Resolution Identification
• Spatial Feature Resolution Identification
• Visual Linearity Perception

36

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Just- Noticeable- Differences (jnd) units
37

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
Icon Brightness l5
(0,0.33)
(0,0.66)
(0,1)
(0.33,0.66)
(0.33,1)
(0.66,1)
38

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
l0.625 l2.5 l5
Icon Spacing
(0,0.33)
(0,0.66)
(0,1)
(0.33,0.66)
(0.33,1)
(0.66,1)
39

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 1 Data Resolution
l0.625 l2.5 l5
Icon Size
(0,0.66)
(0,1)
(0,0.33)
(0.33,0.66)
(0.33,1)
(0.66,1)
40

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
41

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
42

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon Brightness
(0,0.33)
(0,0.66)
(0,1)
(0.66,1)
(0.33,0.66)
(0.33,1)
43

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon Spacing
(0,0.33)
(0,0.66)
(0,1)
(0.33,0.66)
(0.33,1)
(0.66,1)
44

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 2 Spatial Feature Resolution
Icon Size
(0,0.33)
(0,0.66)
(0,1)
(0.33,0.66)
(0.33,1)
(0.66,1)
45

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon spacing
46

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon spacing
47

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Task 3 Visual Linearity
Icon spacing
48

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Valid Experimental Methodology?

• PROS
• Better than expected results
• Simple tasks, yet powerful
• Low variance
• Good initial experiment

• CONS
• Fatigue
• No control of context variables
• Visual linearity failed
• Only single-variable visualizations

49

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Main conclusion

• Successful quantification of perceptual
  interaction (filtering interference) among
  three visual dimensions
• Data resolution (DR)
• Spatial Feature Resolution (FR)

DR
FR
DR
Spacing(0,1)
Brightness(0,1)
Brightness(0.33,0.66)
50

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Effective Visualization Methods

• We can now select an effective visualization
  method

DR6 FR8
DR15 FR4
51

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

DR15 FR4
52

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

DR6 FR8
53

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

54

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

55

• Background
• Exploratory Visualization
• Experiment Details
• Results
• Discussion

Whats next?

• Study saliency
• Weights on requirements
• Include other visual dimensions
• Analyze combinations (two data variables)
  using expert visual designers
• Apply to real datasets

56
Acknowledgements

• NSF ITR grant CNS-0427374
• Experiments subjects
• Paper reviewers
• Graduate students and staff from
• Browns Graphics Group
• Browns Visualization Research Lab
• RISD Illustration professors
• Cullen Jackson, Jason Sobel, and Shiyin Wang

57
Subjective Quantification ofPerceptual
Interactionsamong some 2DScientific
Visualization Methods
Daniel Acevedo daf_at_cs.brown.edu David
Laidlaw dhl_at_cs.brown.edu
http//vis.cs.brown.edu