Got a YouTube account?

New: enable viewer-created translations and captions on your YouTube channel!

English subtitles

← Perception of Visual Cues - Intro to Data Science

Get Embed Code
5 Languages

Showing Revision 5 created 05/25/2016 by Udacity Robot.

  1. In 1985, two scientists from AT&T labs published
  2. a paper on graphical perception and methods. The study
  3. determined how accurately people read the visual cues that
  4. are presented here. This resulted in a ranked list
  5. of most accurate to least accurate visual cues. So
  6. you can see that position is more accurate. Where
  7. as hue is less accurate. What does accuracy mean
  8. here? Well, in this case, it just means how
  9. easily are people able to perceive the values in
  10. your data set given the visual encoding that you've
  11. chosen. You might wonder why saturation and hue are
  12. considered inaccurate. Well, this is a great example. One should
  13. be a little bit cautious when using color, hues
  14. and saturation as visual cues. Like all aspects of
  15. visual perception, we don't perceive color in an absolute
  16. manner. For example, these are all different shades of gray.
  17. When trying to compare and contrast them, it's difficult to
  18. know which shades of gray correspond to exactly which values,
  19. or just how much darker one shade of gray is
  20. than another. For this reason, it's really hard for viewers of
  21. your visualization to really know what a different shade of
  22. gray might mean. How much bigger is one data point
  23. than another? Because of this, you should be careful when
  24. using color, hues and saturation, to encode information in your visualization.
  25. This ranking up here is really a oversimplification of
  26. how visualization works. So you should really use it as
  27. a guide, rather than a definitive rule buck. Efficiency and
  28. exactness are not always the goal of our visualization. Sometimes
  29. color, saturation or color, hue or volume or area can
  30. be really effective to communicate what we are trying to
  31. tell the viewer. However it's good to know generally how
  32. well people will be able to read different visual cues.