Information Visualization Crash Course (AKA Information Visualization 101) Chad Stolper Assistant Professor Southwestern University (graduated from Georgia Tech CS PhD) 1
What is Infovis? Why is it Important? Human Perception Chart Basics (If Time, Some Color Theory) The Shneiderman Mantra Where to Learn More 2
What is Information Visualization? 3
Information Visualization “The use of co compu puter er -supported, in interact eractiv ive , visual representations of abstract data to vi am amplif plify co cognit itio ion .” Card, Mackinlay, and Shneiderman 1999 4
Co Communication Ex Expl ploratory Data Ana nalysis (EDA) 5
Com Communi unication on (gone wrong) 6
7
Edward Tufte An American statistician and professor emeritus of political science, statistics, and computer science at Yale University. He is noted for his writings on information design and as a pioneer in the field of data visualization. -Wikipedia 8 � X
Sp Space Sh Shuttle le C Challe llenger January 28, 1986 Morning Temperature: 31°F 9
10
Tufte, E. R. (2012). Visual explanations: images and quantities, evidence and narrative . Cheshire, CT: Graphics Press. 11
Most Watched Science Experiment Richard Feynman, Physics Nobel laureate explained how rubber became rigid in cold temperate YouTube video: https://youtu.be/6Rwcbsn19c0 Video originally from: http://www.FeynmanPhysicsLectures.com 13
How did this happen? 14
Engineers at Morton Thiokol , the rocket maker, presented on the day before and recommended not to launch. Tufte, E. R. (2012). Visual explanations: images and quantities, evidence and narrative . Cheshire, CT: Graphics Press. 15
19
24
25
26
27
28
So, communication is ortant . ex extr trem emel ely y importan Visualization can help with that – com communicat cate e ideas eas an and insigh ghts . 29
http://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen.html 30
Visualization can also help with Ex Expl ploratory Data Ana nalysis (EDA) But But why hy do do you u ne need d to ex explor ore e dat ata a at at al all??? 31
“There are three kinds of lies: lies, damned lies, and statistics.” 33
Mystery Data Set 34
Mystery Data Set Pr Proper erty Va Value mean( x ) 9 variance ( x ) 11 mean( y ) 7.5 variance ( y ) 4.122 correlation ( x,y ) 0.816 Linear Regression Line y = 3 + 0.5x 35
36
37
38
39
Anscombe’s Quartet https://en.wikipedia.org/wiki/Anscombe%27s_quartet 40
Anscombe’s Quartet Sanity Checking Models Outlier Detection 41
Data visualization leverages hum human n pe percept ption 43
Name the five senses. 44
Sense Sense Ba Bandwidth (bi bits/ s/sec) sec) Sight 10,000,000 Touch 1,000,000 Hearing 100,000 Smell 100,000 Taste 1,000 http://www.britannica.com/EBchecked/topic/287907/information-theory/214958/Physiology 45
A (Simple) Model of Human Visual Perception 46
A (Simple) Model of Human Perception Stage 2 Stage 1 Parallel detection of Serial processing of basic features into object identification and an iconic store spatial layout 47
Stage 1: Pre-Attentive Processing Rapid Parallel Automatic (Fleeting = lasting for a short time) 48
Stage 2: Serial Processing Relatively Slow (Incorporates Memory) Manual 49
Stage 1: Pre-Attentive Processing The eye moves every 200ms (so this processing occurs every 200ms-250ms) 50
Example 1281768756138976546984506985604982826762 9809858458224509856458945098450980943585 9091030209905959595772564675050678904567 8845789809821677654876364908560912949686 51
Example 1281768756138976546984506985604982826762 9809858458224509856458945098450980943585 9091030209905959595772564675050678904567 8845789809821677654876364908560912949686 52
A few more examples from Prof. Chris Healy at NC State 53
Left Side Right Side 54
Raise your hand if a RE RED DOT is present… (On the left or on the right?) 55
56
57
Color (hue) is pre-attentively processed. 58
Raise your hand if a RED DOT is present… 59
60
61
Shape is pre-attentively processed. 62
Determine if a RED DOT is present… 63
64
65
Hue and shape together are NOT pre-attentively processed. 66
Pre-Attentive Processing length hue • • width lightness • • size flicker • • curvature direction of motion • • number binocular lustre • • terminators stereoscopic depth • • intersection 3-D depth cues • • closure lighting direction • • 67
Stephen Few “Now You See It” pg. 39 68
Pre-Attentive à Cognitive 69
Gestalt Psychology Berlin, Early 1900s 70
Gestalt Psychology Goal was to understand pattern perception Gestalt (German) = “seeing the whole picture all at once” instead of a collection of parts Identified 8 “Laws of Grouping” http://study.com/academy/lesson/gestalt-psychology-definition-principles-quiz.html 71
Gestalt Psychology 1. Proximity 7. Good Gestalt 2. Similarity 8. Past Experience 3. Closure 4. Symmetry 5. Common Fate 6. Continuity 72
How many groups are there? 73
74
Proximity 75
How many groups are there? 76
77
Similarity 78
How many shapes are there? 79
80
Closure 81
How many items are there? 82
( ) { } [ ] 83
Symmetry ( ) { } [ ] 84
How many sets are there? 85
86
Common Fate 87
How many objects are there? 88
89
Continuity 90
How many objects are there? 91
92
Good Gestalt 93
What is this word? 94
FLIGHT 95
Past Experience FLIGHT 96
Pre-Attentive Processing Gestalt Laws 101
Detect Quickly 102
Detect quickly does NOT mean detect accurately Id Ideally you want both. 103
Angles Circular Positions areas Rectangular areas (aligned or in a treemap) Crowdsourcing Graphical Perception: Using Mechanical Turk to Assess Visualization Design.Heer 104 and Bostock. Proc ACM Conf. Human Factors in Computing Systems (CHI) 2010, p. 203–212.
Crowdsourced Results T1 T2 T3 Positions T4 T5 T6 Angles T7 Circular areas T8 Rectangular areas T9 (aligned or in a treemap) 1.0 1.5 2.0 2.5 3.0 Log Error Crowdsourcing Graphical Perception: Using Mechanical Turk to Assess Visualization Design.Heer 105 and Bostock. Proc ACM Conf. Human Factors in Computing Systems (CHI) 2010, p. 203–212.
125 Automating the Design of Graphical Presentations l I I Position More accurate IMll 1 I Length F-l Fig. 14. Accuracy ranking of quantitative perceptual tasks. Higher tasks are accom- plished more accurately than lower tasks. Iha I Cleveland and McGill empirically verified the I0.I basic properties of this ranking. I I Volume rl l¶kJ cl Color mot Less accurate Shown) Mackinlay, 1986 Nominal Ordinal 106 Quantitative Position Position Position Color Hue Texture Color Saturation Connection Containment Density Color Saturation Shape Color Saturation Length Angle Slope Area Volume Fig. 15. Ranking of perceptual tasks. The tasks shown in the gray boxes are not relevant to these types of data. An example analysis for area perception is shown in Figure 16. The top line shows that a series of decreasing areas can be used to encode a tenfold quantitative range. Of course, in a real diagram such as Figure 13, the areas would be laid out randomly, making it more difficult to judge the relative sizes of different areas accurately (hence, area is ranked fifth in Figure 14). Nevertheless, small mis- judgments about the size of an area only leads to small misperceptions about the corresponding quantitative value that is encoded. The middle line shows that area can encode three ordinal values. However, one must be careful to make sure ACM Transactions on Graphics, Vol. 5, No. 2, April 1986.
Stephen Few “Now You See It” pg. 41 107
What does this tell us? 108
Barcharts, scatterplots, and line charts are really effective for quantitative data 40 40 20 20 0 0 0 20 40 0 20 40 0 20 40 109
(and for statistical distributions) Tukey Box Plots 113
Outliers Largest < Q3 + 1.5 IQR Largest < Q3 Median Smallest > Q1 Smallest > Q1 - 1.5 IQR 114
115
Tufte’s Chart Principles Edward Tufte 116
Tufte’s Chart Principles DO NOT LIE! Maximize Data-Ink Ratio Minimize Chart Junk 119
Recommend
More recommend
