Tweet
Tweet

Slide 1

Slide 1 text

Data Visualization Science Boot Camp SE
 NCSU Library July 17, 2014 Christopher G. Healey Department of Computer Science North Carolina State University [email protected] http://www.csc.ncsu.edu/faculty/healey
 NC STATE UNIVERSITY sic parvis magna

Slide 2

Slide 2 text

Visualization • Harness viewer’s strengths! • human visual system! • pattern recognition capabilities! • domain expertise! • understanding context! • ability to manage ambiguity! • Manage collaboration between
 viewer and computer! • Enhance each participant’s
 individual strengths! • Share initiative to offset
 their weaknesses Painterly visualization of a slice through a simulated! supernova collapse: pressure → luminance,! velocity → hue, flow direction → orientation! ! Data courtesy Dr. Jon Blondin, Astrophysics, NCSU

Slide 3

Slide 3 text

Painterly Visualization Painterly visualization of a slice through a simulated supernova collapse: pressure → luminance, velocity → hue, flow direction → orientation ! Data courtesy Dr. Jon Blondin, Astrophysics, NCSU Tateosian, Healey, Enns. “Engaging Viewers Through Nonphotorealistic Visualizations,” NPAR 2007, 93–102, 2007.

Slide 4

Slide 4 text

Graphs • Traditional method for comparing data attributes! • E.g., line chart, bar chart, scatterplot, histogram
 • Different graphs are used for different tasks:! • tracking changes (line graph)! • correlations, relationships (X-Y plot)! • correlations, clustering (scatterplot)! • category comparison (bar chart)! • distributions (pie chart)! • uniqueness and overlap (Venn diagram)
 • Graph elements can represent multiple attribute values

Slide 5

Slide 5 text

Line Chart U.S. Unemployment Line Chart ! U3: people without employment, actively searching in the last 4 weeks U6: U3 + people not looking + people who gave up + part-time workers who want full-time jobs

Slide 6

Slide 6 text

Line Chart U.S. Unemployment By Recession Line Chart

Slide 7

Slide 7 text

Stacked Line Chart Stacked time series line chart of baby name popularity from 1900–2003 ! Data Courtesy

Slide 8

Slide 8 text

Bar Chart May 2013 Unemployment By State Bar Chart

Slide 9

Slide 9 text

Pie Chart Exploded 3d pie chart of transportation use categorized by transportation type

Slide 10

Slide 10 text

Pie Chart Pie chart of population by U.S. state

Slide 11

Slide 11 text

Histogram Interval histogram (with overlay for comparison) of pixel brightnesses binned by greyscale

Slide 12

Slide 12 text

Scatterplot Petal Length Petal Length Sepal Width class (Number of sepals on a flower is its merosity)

Slide 13

Slide 13 text

Maps • Traditional method for presenting geographic and geospatial data! • E.g., topographic, planimetric, base
 • Different maps are used for different
 tasks:! • choropleth (charting value by region)! • isarithmic (contour lines)! • proportional symbol (point location data)! • dot (presence or absence of a feature)
 • Maps can be extended to visualize
 multidimensional data Topographic map from Indiana Jones and the Last Crusade

Slide 14

Slide 14 text

Choropleth Map 2013 Percentage of U.S. households in poverty by county

Slide 15

Slide 15 text

Choropleth Map Classification of U.S. counties into twelve “groups” ! Data courtesy

Slide 16

Slide 16 text

Isarithmic Map Thickness of the Earth’s crust, 10km intervals ! Data courtesy

Slide 17

Slide 17 text

Proportional Symbol Map Number of venture capital deals in 2012 ! Data courtesy ghost-towns-of-the-new-economy/309460/

Slide 18

Slide 18 text

Dot Map John Snow’s cholera map ! Data courtesy

Slide 19

Slide 19 text

Route Map Henry Beck’s London Underground map circa 1933 ! Data courtesy

Slide 20

Slide 20 text

Route Map London Underground map circa 2012 ! Data courtesy

Slide 21

Slide 21 text

Napolean’s
 March on Russia Charles Joseph minard, 1869, flow map of napolean’s march on russia

Slide 22

Slide 22 text

“Preattentive” Features • Certain basic visual features are detected by our low-level visual system! • detection is rapid, usually in one “glance” of 100–250 msec! • can determine presence or absence, possibly amount! • unique features can capture our focus of attention
 • Initially proposed as an automatic, bottom-up phenomena! • Treisman’s feature map theory, feature hierarchies
 • Combined bottom-up and top-down models also exist! • Wolfe’s guided search, Huang et. al’s boolean maps

Slide 23

Slide 23 text

Visual Interference • Design, conduct, and analyze target detection experiments! • Study colour: luminance, hue; and texture: size, orientation! • Effectiveness of visual features in isolation, pairwise interference 0.06° 0.1225° 0.245° Visual acuity experiment trials — 0.06°, 0.1225°, 0.245° — with novel target detection hue target size target orientation target

Slide 24

Slide 24 text

• Design, conduct, and analyze target detection experiments! • Study colour: luminance, hue; and texture: size, orientation! • Effectiveness of visual features in isolation, pairwise interference Visual Interference 0.06° 0.1225° 0.245° Visual acuity experiment trials — 0.06°, 0.1225°, 0.245° — with novel target detection hue target size target orientation target w/random hue Healey and Sawant. “On the Limits of Resolution and Visual Angle in Visualization,” ACM TAP 9 (4), 2012.

Slide 25

Slide 25 text

Postattentive Amnesia • Does previewing make us faster?! • Intuition suggests it will! • Extract detail! • Access it rapidly on demand! ! • Experiments show that human
 vision does not work this way! • Vision is not a camera! • Detail is only available at the most
 recent focus of attention priming image

Slide 26

Slide 26 text

Postattentive Amnesia • Does previewing make us faster?! • Intuition suggests it will! • Extract detail! • Access it rapidly on demand! ! • Experiments show that human
 vision does not work this way! • Vision is not a camera! • Detail is only available at the most
 recent focus of attention priming image Purple Tilted Wolfe, Klempen, Dahlen, “Post Attentive Vision,” JEP: HPP 26 (2), 2000.

Slide 27

Slide 27 text

Primed Search Green Vertical

Slide 28

Slide 28 text

Primed Search Green Vertical

Slide 29

Slide 29 text

Primed Search

Slide 30

Slide 30 text

Primed Search White Tilted

Slide 31

Slide 31 text

Find Five Differences

Slide 32

Slide 32 text

Find Five Differences eyes tilted up bee’s stripe colours reversed extra leaf patch on knee extra flower

Slide 33

Slide 33 text

Change Blindness Data courtesy Dr. Ron Rensink, Department of Psychology, UBC

Slide 34

Slide 34 text

Change Blindness

Slide 35

Slide 35 text

Change Blindness Rensink. “Seeing, Sensing, and Scrutinizing,” Vision Research 40, (10-12), 2000.

Slide 36

Slide 36 text

Change Blindness Models • Overwriting! • previous image overwritten! • First impression! • initial view abstracted! • Nothing is stored! • scene abstracted with no detail! • Feature combination! • previous and new views
 combined! • Everything is stored, nothing is compared! • details cannot be accessed without external stimulus Main actor changes across movie cut Simons. “Current Approaches to Change Blindness,” Visual Cognition 7 (1-3), 2000.

Slide 37

Slide 37 text

Determine Arm Length

Slide 38

Slide 38 text

Determine Arm Length

Slide 39

Slide 39 text

Inattentional Blindness grey square in frame not detected during attention-demanding task Mack and Rock. Inattentional Blindness. MIT Press, 2000.

Slide 40

Slide 40 text

Inattentional Blindness • Attention-demanding task renders viewers “blind” to change! • Change is easily seen if scene is viewed without any task
 • Limited “visual attention”! • Classic example, counting basketball passes! • Woman wearing a gorilla suit walks through the scene! • More than half of the viewers do
 not notice the gorilla Superimposed video streams Single video stream Woman with umbrella Woman with umbrella Gorilla Gorilla Simons and Chabris, “Gorillas in Our Midst: Sustained Inattentional Blindness for Dynamic Events,” Perception 28 (9), 1999.

Slide 41

Slide 41 text

Nonphotorealism • Visual system operates in two stages! • orientation: focus of attention snaps to a location in an image! • engagement: visual system chooses to linger at the location, taking in detail
 • Local feature differences orient! • Hypothesize that increased visual aesthetic produces engagement! • Measure engagement using memory for detail indication & detail painting style visual complexity painting style

Slide 42

Slide 42 text

Interpretational Complexity (IC) Painterly Styles Indication & Detail (ID) Visual Complexity (VC) < < Tateosian, Healey, Enns. "Engaging Viewers Through Nonphotorealistic Visualizations,” NPAR 2007, 93–102, 2007.

Slide 43

Slide 43 text

Memorization Observe each of the following images, and commit them to memory

Slide 44

Slide 44 text

Memorization Indication & Detail (ID) Painting Style

Slide 45

Slide 45 text

Memorization

Slide 46

Slide 46 text

Memorization Interpretational Complexity (IC) Painting Style

Slide 47

Slide 47 text

Memorization

Slide 48

Slide 48 text

Memorization Visual Complexity (VC) Painting Style

Slide 49

Slide 49 text

Memorization

Slide 50

Slide 50 text

Memorization Traditional Glyph Visualization

Slide 51

Slide 51 text

No content

Slide 52

Slide 52 text

Recall For each of the following images, determine whether it matches one of the images you memorized ! Answer “Yes” if it does ! Answer “no” if it does not

Slide 53

Slide 53 text

Recall

Slide 54

Slide 54 text

Recall No

Slide 55

Slide 55 text

Recall

Slide 56

Slide 56 text

Recall Yes

Slide 57

Slide 57 text

Recall

Slide 58

Slide 58 text

Recall Yes

Slide 59

Slide 59 text

Recall

Slide 60

Slide 60 text

Recall No

Slide 61

Slide 61 text

Recall

Slide 62

Slide 62 text

Recall No

Slide 63

Slide 63 text

Recall

Slide 64

Slide 64 text

Recall No

Slide 65

Slide 65 text

Memorization

Slide 66

Slide 66 text

Recall Yes

Slide 67

Slide 67 text

Recall

Slide 68

Slide 68 text

Recall Yes

Slide 69

Slide 69 text

Engagement • Perceived aesthetic can be systematically varied within a visualization! ! • Increasing aesthetic improves memory for detail! • Viewers had recall accuracy of 80% or more, even 24 hours after exposure! ! • Accuracy for most aesthetic visualization no better than for traditional visualization! • Traditional visualizations are seen as aesthetic?! • More detailed models of engagement needed?! • Memory for detail is not an appropriate measure of engagement?

Slide 70

Slide 70 text

Contact Information NC STATE UNIVERSITY [email protected] http://www.csc.ncsu.edu/faculty/healey ! ! 
 Special Thanks to: John Blondin (Astrophysics, NCSU)
 James Enns (Psychology, UBC)
 Ron Rensink (CS & Psychology, UBC) ! Geniva Liu, Mark Remple,
 Amit Sawant, Laura Tateosian