1. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 0 Displaying Quantitative Information An exploration of Edward R. Tufte’s The Visual Display of Quantitative Information Jeffrey Nichols Programming Usable Interfaces May 2, 2003

2. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 1 Good graphics Napolean’s invasion of Russia, as drawn by Charles Joseph Minard (1781-1870)

3. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 2 More good graphics The price of wheat compared to labour wages, William Playfair (1759-1823)

4. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 3 More good graphics French train schedule, as drawn by E.J. Marey (1830-1904)

5. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 4 More good graphics Map of the northern galactic hemisphere (1.3 million galaxies shown)

6. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 5 What made these graphics good?

7. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 6 Bad graphics

8. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 7 More bad graphics

9. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 8 More bad graphics

10. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 9 More bad graphics

11. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 10 More bad USA Today Graphs

12. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 11 What made these graphics bad?

13. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 12 How can we make better graphics? Tufte presents some principles of data graphics  Above all else, show the data.  Maximize the data-ink ratio  Erase non-data-ink  Erase redundant data-ink  Revise and edit

14. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 13 Above all else, show the data

15. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 14 Show the data

16. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 15 Data-Ink  Ink that changes as the data changes  Non-redundant ink!

17. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 16 Data-Ink Ratio Data-ink ratio = data-ink Total ink used to print graphic =Proportion of a graphic’s ink devoted to the non-redundant display of data-information. =1.0 – proportion of graphic that can be erased without the loss of information

18. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 17 Ratio of Histogram What is the data-ink ratio of this graphic? < 0.05 !!!

19. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 18 Ratio of USA Today Graphic  What is the data-ink ratio of this graphic? < 0.001

20. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 19 How can we make better graphics? Tufte presents some principles of data graphics  Above all else, show the data.  Maximize the data-ink ratio  Within reason  Every bit of ink on a graphic requires a reason  Erase non-data-ink  Erase redundant data-ink  Revise and edit

21. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 20 How can we make better graphics? Tufte presents some principles of data graphics  Above all else, show the data.  Maximize the data-ink ratio  Erase non-data-ink  Erase redundant data-ink  Revise and edit

22. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 21 Erase non-data-ink, within reason

23. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 22 Erase non-data-ink, within reason

24. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 23 Erase non-data-ink, within reason Histogram of Midterm Results Scoring Buckets # of Students C B- B B+ A- A 0 1 2 5 11 18 7

25. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 24 Erase redundant data-ink Histogram of Midterm Results Scoring Buckets # of Students C B- B B+ A- A 0 1 2 5 11 18 7

26. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 25 Erase redundant data-ink Histogram of Midterm Results Scoring Buckets # of Students C B- B B+ A- A 0 1 2 5 11 18 7 What’s the data-ink ratio?

27. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 26 How can we make better graphics? Tufte presents some principles of data graphics  Above all else, show the data.  Maximize the data-ink ratio  Erase non-data-ink  Erase redundant data-ink  Revise and edit

28. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 27 The Principles Applied to Common Forms Box Plots Median 75% 25% Maximum Minimum

29. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 28 The Principles Applied to Common Forms Box Plots Median 75% 25% Maximum Minimum

30. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 29 The Principles Applied to Common Forms Bar Graphs Original Graph Modified Graph

31. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 30 The Principles Applied to Common Forms Scatter Plots Original Graph Modified Graph

32. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 31 The Principles Applied to Common Forms Another Scatter Plot Variant

33. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 32 Conclusions  Show data variation, not design variation  Avoid using ink for non-data items  Avoid redundancy  Clear and detailed labeling should be used to defeat graphical distortion  Revise and Edit

34. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 33 Questions?

35. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 34 Thanks for a great class! Reminder: Final Exam on Friday, May 9 th at 8:30am in this room

36. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 35 More Bad USA Today Graphs

37. Jeffrey Nichols Displaying Quantitative Information May 2, 2003 Slide 36 Tufte’s Principles  Graphical Integrity  The representation of numbers, as physically measured on the surface of the graphic itself, should be directly proportional to the numerical quantities represented.  Clear, detailed, and thorough labeling should be used to defeat graphical distortion and ambiguity. Write out explanations of the data on the graphic itself. Label important events in the data.  Show data variation, not design variation.  In time-series displays of money, deflated and standardized units of monetary measurement are nearly always better than nominal units.  The number of information-carrying (variable) dimensions depicted should not exceed the number of dimensions in the data.  Graphics must not quote the data out of context.  Theory of Data Graphics  Above all else, show the data.  Maximize the data-ink ratio  Erase non-data-ink  Erase redundant data-ink  Revise and edit  Other comments  Graphical elegance is often found in simplicity of design and complexity of data  Data graphics are paragraphs about data and should be treated as such