Visualization

A talk about the theory of visualization

Transcript

1. By Eitan Lees Visualization
3. “The ubiquity of visual metaphors in describing cognitive processes hints at a nexus of relationships between what we see and what we think” - Mackinlay & Card (1999)
4. External Cognition
7. Part 1: Data Wrangling Part 2: Visual Encodings Part 3: Graphical Critique Part 4: Practical Advice
8. Part 1: Data Wrangling
9. Clean data sets are all alike; every unclean data set is unclean in its own way
10. Things to Consider: - Make Numbers ⇒ Numbers - Make Dates ⇒ Dates - Make Nans ⇒ Nans - Make sure strings aren’t corrupted
11. Before we visualize, let’s tidy up
12. country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583
13. country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 Variables
14. country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 Variables country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 Observations
15. country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583
16. country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583 Variables country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583
17. country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583 Variables country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583 country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583 Observations
18. country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583
19. country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583 We want to gather the values corresponding to each key.
20. country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 country year key value Afghanistan 1999 cases 745 Afghanistan 1999 population 19987071 Afghanistan 2000 cases 2666 Afghanistan 2000 population 20595360 Brazil 1999 cases 37737 Brazil 1999 population 172006362 Brazil 2000 cases 80488 Brazil 2000 population 174504898 China 1999 cases 212258 China 1999 population 1272915272 China 2000 cases 213766 China 2000 population 1280428583 Tidy We want to gather the values corresponding to each key.
21. country 1999 2000 Afghanistan 745 2666 Brazil 37737 80488 China 212258 213766
22. country 1999 2000 Afghanistan 745 2666 Brazil 37737 80488 China 212258 213766 We want to spread the values to the corresponding keys.
23. country year cases Afghanistan 1999 745 Afghanistan 2000 2666 Brazil 1999 37737 Brazil 2000 80488 China 1999 212258 China 2000 213766 country 1999 2000 Afghanistan 745 2666 Brazil 37737 80488 China 212258 213766 We want to spread the values to the corresponding keys.
24. country year cases Afghanistan 1999 745 Afghanistan 2000 2666 Brazil 1999 37737 Brazil 2000 80488 China 1999 212258 China 2000 213766 country 1999 2000 Afghanistan 745 2666 Brazil 37737 80488 China 212258 213766 Tidy We want to spread the values to the corresponding keys.
25. Tidy Data country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 Variables country year cases population Afghanistan 1999 745 19987071 Afghanistan 2000 2666 20595360 Brazil 1999 37737 172006362 Brazil 2000 80488 174504898 China 1999 212258 1272915272 China 2000 213766 1280428583 Observations Values
26. Part 2: Visual Encoding
27. Compare area of circles
28. Compare length of bars
29. Compare length of bars
30. Compare area of circles
31. Length Area Slope Position Angle Volume Color Value Color Hue Shape Visual Encoding Channels And many more ...
32. Length Area Slope Position Angle Volume Color Value Color Hue Shape Accuracy ranking of quantitative perceptual tasks. Better Worse
33. Data Models
34. Nominal: - Labels and Categories - Example: Pill Shape - Operations: =, ≠
35. Nominal: - Labels and Categories - Example: Pill Shape - Operations: =, ≠ Ordinal: - Ordered Sets - Example: Drug Schedule - Operations: =, ≠, <, >
36. Nominal: - Labels and Categories - Example: Pill Shape - Operations: =, ≠ Ordinal: - Ordered Sets - Example: Drug Schedule - Operations: =, ≠, <, > Quantitative: - Numerical Measurement - Example: Dosage - Operations: =, ≠, <, >, -, %
39. 2D Plane Size Color Value Texture Color Hue Angle Shape
40. 2D Plane Size Color Value Texture Color Hue Angle Shape Suitable for Ordered Data Suitable for Unordered Data
41. 2D Plane Size Color Value Texture Color Hue Angle Shape Suitable for Ordered Data Position Area Color Value
42. 2D Plane Size Color Value Texture Color Hue Angle Shape Suitable for Unordered Data Angle Color Hue Shape
43. Position N O Q Size N O Q Color Value N O Q Texture N O Color Hue N Angle N Shape N Nominal Ordinal Quantitative Note: Q⊂O⊂N Bertin’s Levels of Organization
45. Grammar of Graphics 1. Data 2. Transformations 3. Marks 4. Encoding - mapping from fields to mark properties 5. Scale - functions that map data to visual scales 6. Guides - visualizations of scales (axes, legends, etc.)
46. Building Blocks of Visualization
47. Part 3: Graphical Critique
48. Most of modern statistical graphics can be traced back to William Playfair a Scottish engineer and political economist. William Playfair
52. Measle cases per 100000 people
53. Connectivity diagram of character in Les Misérables
54. Data Visualization is Everywhere!
55. Edward Tufte
56. Edward Tufte “Graphical excellence is that which gives to the viewer the greatest number of ideas in the shortest time with the least ink in the smallest space.” ― Edward R. Tufte, The Visual Display of Quantitative Information
57. Edward Tufte “Graphical excellence is that which gives to the viewer the greatest number of ideas in the shortest time with the least ink in the smallest space.” ― Edward R. Tufte, The Visual Display of Quantitative Information
58. Edward Tufte “Graphical excellence is that which gives to the viewer the greatest number of ideas in the shortest time with the least ink in the smallest space.” ― Edward R. Tufte, The Visual Display of Quantitative Information (within reason!)
59. Data to Ink Ratio
60. Data to Ink Ratio
61. Data to Ink Ratio Reasonable?
62. 75% 50% 25% min max Tukey Data to Ink Ratio
63. Tukey Data to Ink Ratio
64. Tukey Tufte #1 Data to Ink Ratio
65. Tukey Tufte #1 Tufte #2 Data to Ink Ratio
66. Tukey Tufte #1 Tufte #2 Data to Ink Ratio Unreasonable?
67. Sparklines “A sparkline is a small intense, simple, word-sized graphic with typographic resolution … ” - Edward Tufte, Beautiful Evidence, p. 46-63.
68. Small Multiples “At the heart of quantitative reasoning is a single question: Compared to what? Small multiple designs answer directly by visually enforcing comparisons of changes, of the differences among objects, of the scope of alternatives.” - Edward Tufte, Envisioning Information, p. 67
71. Yeah, well, that's just, like, your opinion, man.
72. Part 4: Practical Advice
73. Ten Simple Rules for Better Figures By Nicolas P. Rougier 1. Know your audience 2. Identify your message 3. Adapt the figure to the support medium 4. Captions are not optional 5. Do not trust the defaults 6. Use color effectively 7. Do not mislead the reader 8. Avoid “Chartjunk” 9. Message trumps beauty 10. Get the right tool
74. 1. Know your audience 2. Identify your message 3. Adapt the figure to the support medium 4. Captions are not optional 5. Do not trust the defaults 6. Use color effectively 7. Do not mislead the reader 8. Avoid “Chartjunk” 9. Message trumps beauty 10. Get the right tool Ten Simple Rules for Better Figures By Nicolas P. Rougier
75. 1. Know your audience 2. Identify your message 3. Adapt the figure to the support medium 4. Captions are not optional 5. Do not trust the defaults 6. Use color effectively 7. Do not mislead the reader 8. Avoid “Chartjunk” 9. Message trumps beauty 10. Get the right tool Ten Simple Rules for Better Figures By Nicolas P. Rougier
76. 1. Know your audience 2. Identify your message 3. Adapt the figure to the support medium 4. Captions are not optional 5. Do not trust the defaults 6. Use color effectively 7. Do not mislead the reader 8. Avoid “Chartjunk” 9. Message trumps beauty 10. Get the right tool Ten Simple Rules for Better Figures By Nicolas P. Rougier
77. 1. Know your audience 2. Identify your message 3. Adapt the figure to the support medium 4. Captions are not optional 5. Do not trust the defaults 6. Use color effectively 7. Do not mislead the reader 8. Avoid “Chartjunk” 9. Message trumps beauty 10. Get the right tool Ten Simple Rules for Better Figures By Nicolas P. Rougier