User Tools

Site Tools


bad_map_examples

Visual Analytics Course

Perception

This section criticize bad or not so bad maps or interactive visualizations.

It is often worth looking and learn from bad design, as the Cartastrophe website suggests.

Death Penalty and Chernoff Faces

The designer of this one definitely should have to pay for its bad design. I got this one from the Cartastrophe website. Typing in “Death Penalty Chernoff Face” in your favorite search engine will easily lead you to the map.

The interest aspect of the map is that is wonderfully illustrate the point we are stressing here. That the communication channel can be disrupted at three different levels: operational (perception), articulatory and/or semantic. Each of these level here is undermined by this example.

Operational distance (perception). The choice of colors definitely is not optimal. The main reason is because of poor contrast between colors, but also poor luminance contrast as well. Looking at their RGB values, the three colors used for the map regions are (close to):

RGB coding HSL coding
Purple 134.45.134 Hue 300, Saturation 49.7, Lightness 35.1
Maroon 134.45.45 Hue 0, Saturation 49.7, Lightness 35.1
Green 45.134.45 Hue 120, Saturation 49.7, Lightness 35.1

The poor contrast we observe may thus result from the fact that, taken pairwise, the RGB values of colors coincide on one of the components. But more convincingly, the HSL coding clearly shows why the colors show bad contrast: they have equal saturation and luminosity.

Articulatory distance. This is the level at which the map performs worst.

A first observation is that although the map describes how death penalty conditions and statistics distribute over the USA, the map need not be a faithful geographical representation of the country. There is no reason why the user needs a distance scale. This would allow for a light distortion enabling to draw Alaska at a more convenient place and not in its own independent frame.

Second, the map makes use of several legend to convey information.

  • The number of executed prisoners is indicated using ellipses going from horizontally flat ellipses to thiner but vertically expanded ones. Mapping a one-dimensional parameter onto a two-parameter visual cues is misleading and makes it harder to interpret the map content.
  • Moreover, the figures do not increase along a continuous domain but rather are aggregated into classes.
  • The Chernoff faces are enriched with numerous visual cues in order to carry different meanings. Eyes may be round or square, closed or open, x'ed or not, etc. Not only are the faces quite complex but the legend is hard to read.
  • The color code used to convey with or without death penalty states does not properly convey the meaning. A sober solution would have been to simply leave them blank.

Semantic distance. The map also uses the face to convey information about the average age of prisoners that are put to death. It seems that it’s sad when people over 50 are executed, but it’s pretty happy when people under 40 are (!). This obviously leads to unwanted interpretation or at least will rub most people the wrong way.


Back to the Perception section

/net/html/perso/melancon/Visual_Analytics_Course/data/pages/bad_map_examples.txt · Last modified: 2013/10/02 18:50 by melancon