Everything* you need to know about the psychology of perception† to make effective visualisations‡
Hadley Wickham, Rice University
Thursday, July 19, 12
Everything * you need to know about the psychology of perception to - - PowerPoint PPT Presentation
Everything * you need to know about the psychology of perception to make effective visualisations Hadley Wickham, Rice University Thursday, July 19, 12 Everything * you need to know about the psychology of perception to make
Everything* you need to know about the psychology of perception† to make effective visualisations‡
Hadley Wickham, Rice University
Thursday, July 19, 12Everything* you need to know about the psychology of perception† to make effective visualisations‡
* that I can cram into 40 minutes
Hadley Wickham, Rice University
Thursday, July 19, 12Everything* you need to know about the psychology of perception† to make effective visualisations‡
* that I can cram into 40 minutes
† but I’m not a psychologist
Hadley Wickham, Rice University
Thursday, July 19, 12Everything* you need to know about the psychology of perception† to make effective visualisations‡
‡ in my opinion
* that I can cram into 40 minutes
† but I’m not a psychologist
Hadley Wickham, Rice University
Thursday, July 19, 12Match perceptual and data topology
Which represents the larger value?
Thursday, July 19, 12Which represents the larger value?
Thursday, July 19, 12Which represents the larger value?
Thursday, July 19, 12Which represents the larger value?
Thursday, July 19, 12Which represents the larger value?
Thursday, July 19, 12Which represents the larger value?
Thursday, July 19, 12Which represents the larger value?
Thursday, July 19, 12Which represents the larger value?
Thursday, July 19, 12What are the three important components
r g b
+50/255 R +50/255 G +50/255 B +100/255 R +100/255 G +100/255 B Thursday, July 19, 12(aka polar LUV)
Thursday, July 19, 12hcl hsv hsl
http://en.wikipedia.org/wiki/HSV_color_space#Disadvantages Thursday, July 19, 12(aka polar LUV)
Thursday, July 19, 12h c l
Thursday, July 19, 12chroma
luminance
hue
Munsell, Albert H. (1905). A Color Notation. Boston: G. H. Ellis Co. Thursday, July 19, 12Why care?
Perceptually uniform Hue is unordered. Use evenly spaced hues with equal chroma and luminance to make aesthetically pleasing discrete palettes. Chroma and luminance are ordered. Easy to make perceptually uniform gradients by varying either (or both). Never use rainbow scales again!
Thursday, July 19, 12Aesthetic Topology Position Ordered Size Ordered Luminance Ordered Chroma Ordered Shape Unordered Hue Unordered
Thursday, July 19, 12Make important comparisons easy
Theory, experimentation and application to the development of graphical methods. Journal of the American Statistical Association, 79 (387):531–554, 1984.
Thursday, July 19, 12We perceive relative differences
Thursday, July 19, 12We perceive relative differences
Thursday, July 19, 12Position Length / Angle Area Volume / Chroma / Luminance Close objects are easier to compare than distant objects × × Perception is relative
Thursday, July 19, 12Common misunderstanding Pie charts are bad! Die pie chart, DIE
Thursday, July 19, 12Common misunderstanding Pie charts are bad! Die pie chart, DIE Pie charts are bad when you want to accurately compare two numbers
Thursday, July 19, 12Common misunderstanding Pie charts are bad! Die pie chart, DIE Pie charts are bad when you want to accurately compare two numbers But: As good as bars for estimating percentage of whole. Better than bars for comparing compound proportions (A + B vs C + D)
Thursday, July 19, 12Common misunderstanding Pie charts are bad! Die pie chart, DIE Pie charts are bad when you want to accurately compare two numbers But: As good as bars for estimating percentage of whole. Better than bars for comparing compound proportions (A + B vs C + D)
Statistical Chart. Journal of Educational and Behavioral Statistics, 30:353–368, 2005.
Thursday, July 19, 12Beware
Thursday, July 19, 12Beware
http://www.michaelbach.de/ot/sze_sineIllusion/index.html Thursday, July 19, 12Beware
Thursday, July 19, 12Beware
Thursday, July 19, 12Beware
Thursday, July 19, 12Ensure important comparisons are close Use position, then length/area, then chroma/luminance If possible, display comparisons directly
Thursday, July 19, 12Visual connections should reflect real connections
Beware
Thursday, July 19, 12Beware
Thursday, July 19, 12Beware of animation!
Beware of animation!
(Compare in space, not time)
Thursday, July 19, 12We often miss gradual changes too We often don’t notice abrupt changes
http://youtu.be/FWSxSQsspiQ?t=0m12s http://youtu.be/1nL5ulsWMYcAnd movement makes us miss other changes
http://visionlab.harvard.edu/silencing/ Thursday, July 19, 12Match perceptual and data topology
Make important comparisons easy
Visual connections should reflect real connections
Beware of animation!
Visualisation is only
analysis
Transform Visualise Model
Thursday, July 19, 12