The Visual Display of Quantitative Information
By Edward Tufte.
I quickly read this book for the second time recently in order to take notes on it. My notes are below.
- Dr John Snow’s cholera map of 1854 (pg 24)
- Charles Joseph Minard’s map of French wine exports in 1864 (pg 25)
- population distribution usa, 1970 (pg 156)
- Don’t waste on simple data, use numbers for that. Use for highly variable data.
- NYC weather for 1980 (pg 30)
- E. J. Marey’s graphical train schedule for Paris to Lyon in the 1880s. (pg 31)
- J. H. Lambert (early 1700s) and William Playfair (last 1700s) were the great inventors of modern graphical design. (pg 32)
- Minard’s fate of Napoleon’s army in Russia, drawn 1861 (pg 41)
- Thermal conductivity of copper, 1974 (pg 49)
Experiments have shown that we don’t perceive measured data accurately.
- e.g. perceived area of circle grows more slowly than actual:
- reported perceived = (actual area)^x, x=.8+-.3 (pg 55)
- plus you can tell that everyone is different
Representations of 3d objects lead to confusing comparisons as perspective reduces our ability to accurately compare measurements.
- “Representation of numbers should be directly proportional to the numerical quantities represented.”
- “Show data variation, not design variation.” (pg 61)
- “Number of … variables depicted should not exceed the number of dimensions in the data.” (pg 71)
- “Graphics must not quote data out of context.” (pg 74)
No one can write decently who is distrustful of the reader’s intelligence, or whose attitude is patronizing. – E. B. White
- Data-ink ratio = data-ink / total ink used to print graphic
- erase non-data-ink and redundant data-ink
- (it’s only a single dimension to measure with, efficiency, should also consider complexity, structure, density, and beauty. pg 137)
Chartjunk to erase:
- unintentional optical art (patterned shading.”moire vibration”)
- grid lines (can be faint)
- pg 118: perhaps the worst graphic ever, only 5 pieces of data
- the duck (decorative forms)
- bar plot pg 128
- boxplot pg 129
- scatter plot pg 133
Data can be part of graphic, numbers as points or labels as markers for max/min or points values.
- data density = numbers of entries in data matrix/area of data graphic
- e.g. annual sunshine report in London 1971 (pg 165)
Graphs can often be shrunk.
“Graphical elegance is often found in simplicity of design and complexity of data.” (pg 177)