If your Results are Good, Publish; if they are weak, use graphics

Posted by – August 10, 2009

Or so joked my MS Advisor, Jim Warner.

Scientists and engineers have a well-deserved reputation for presentations and graphics that obscure their results rather than display them. Reports by the National Research Council on decision support for climate science and policy and by NASA’s Return to Flight Task Group suggest that we need to improve our communication of research results.

In his book, The Visual Display of Quantitative Information, Edward Tufte emphasizes that scientific graphics should communicate complex ideas with clarity, precision, and efficiency. Tufte asserts that:

What is to be sought in designs for the display of information is the clear portrayal of complexity. Not the complication of the simple; rather the task of the designer is to give visual access to the subtle and the difficult…

For him, excellent graphics should:

  • show the data
  • induce reader to think about the data
  • avoid distorting the data
  • put many numbers in small space
  • make large data sets coherent
  • encourage the eyes to compare data measures
  • reveal the data at several levels
  • serve a clear purpose (description, exploration, tabulation, decoration)
  • integrate the data with the text

Tufte provides many examples of graphical excellence throughout history, exhorting authors to maximize the ratio of the data displayed to the ink used in making the chart. He also provides examples of misleading graphs, of charts that are all design and no information, and of chartjunk, the useless decoration of charts that distracts and obscures the data. As examples, Tufte redesigns several commonly used graphics, including the bar chart, the box-and-whisker plot, and the scatterplot.

A book by Tufte is available on-line, as is his essay on the evils of PowerPoint, and an extensive blog.
-ddw

4 Comments on If your Results are Good, Publish; if they are weak, use graphics

  1. [...] how spending is divided, but also burrow down into each agency for additional detail. I think the Tufte perspective would be that the departmental/agency logos don’t tell the reader much – except that [...]

  2. [...] I would argue that the use of bottles as symbols is what Tufte would call chartjunk, this is still a wonderful tool for exploring water supply and water use by [...]

  3. [...] total, and makes good use of the mouse roll-over to reveal the actual budget numbers – very Tufte-friendly. I would suggest, however, that it could be improved by adding links within the mouse roll-over for [...]

  4. [...] of this blog will note that I am an unabashed fan of Edward Tufte, the Da Vinci of graphical analysis. In a recent posting at the Washington Post’s On [...]

Leave a Reply