LIS 504 - Graphic displays of data
Much of this information is derived from
- Tufte, E.R. 1983.
The Visual display of quantitative information.
Cheshire, Connecticut: Graphics Press.
(IMSSTK HA31.T83.)
What graphic displays should do
- Show the data.
- Induce the viewer to think about the substance
rather than something else
(keep the data-ink ratio high).
- Avoid distorting what the data have to say
(graphical integrity).
- Present many numbers in a small space
(high data density).
- Make large data sets coherent.
- Encourage comparison of different pieces of data.
- Reveal the data at different levels of detail.
- Serve a reasonably clear purpose.
- Be closely integrated
with statistical and verbal descriptions.
- Representation of numbers as areas on the graph
should be directly proportional to the numerical quantities represented
(i.e., the lie factor
should be near 1).
- Labels should be clear, detailed and thorough.
- Design variation should reflect data variation.
- Units of measurement should be standardized.
- The number of variable dimensions depicted should not exceed
the number of variables in the data.
- Data should not be quoted out of context.
Lie factor
| lie factor =
| size of effect shown in graph
size of effect in data
|
where
| size of effect =
| |second value - first value|
first value
|
A lie factor that is either much higher or much lower than one is bad.
A high lie factor exaggerates differences between values.
A low lie factor obscures differences between values.
A common example of a high lie factor
occurs when both dimensions of a two-dimensional figure
are made proportional to the same data,
so that the size of the figure is proportional to the square
of the data;
for instance,
| Year
| Books circulated
|
| 2001
| 100
&
|
| 2002
| 141
&
|
| 2003
| 200
&
|
where the lie factor is about 2.4.
An example of a low lie factor
can be seen in the "Cones"
custom chart format in Microsoft Excel.
The heights of the (truncated) cones are proportional to the data,
but their areas on the screen and their apparent volumes
make the larger data values seem relatively small.
Charting on a logarithmic scale can also produce
a low lie factor.
Data-ink ratio
| data-ink ratio =
| "ink" used to display the data
total "ink" used to display the graphic
|
The data-ink ratio should be relatively high.
A low data-ink ratio can be a result of "chartjunk".
Types of chartjunk include
- unnecessary and distracting patterns
("moiré vibration"),
- obtrusive grids, and
- graphics emphasizing style over quantitative information
("ducks")
Data density
| data density =
| number of entries in data matrix
area of data graphic
|
The data density should be relatively high.
If there are not enough data points to support a high data density
in a graphic,
using a simple table may be a better idea.
Home
Last updated November 6, 2000.
This page maintained by
Prof. Tim Craven
E-mail (text/plain only): t.craven@uwo.ca
Faculty of Information and Media Studies
University of Western
Ontario,
London, Ontario
Canada, N6A 5B7