Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: st: how to graph pie 3d-solid?
Nick Cox <email@example.com>
RE: st: how to graph pie 3d-solid?
Sun, 17 Oct 2010 18:02:03 +0100
Maarten's answers are on two levels, briefly that you can't do it in Stata and at greater length that it's a very bad idea any way.
To add to both a little, and freely adding personal opinions:
1. In principle this is programmable, but it does not look trivial.
2. Pie charts are usually greatly inferior to bar charts or dot charts, although their familiarity and the ease with which fractions close to 1/2 or 1/4 can easily be decoded make them not quite worthless. But pie charts with a false third dimension are much worse. At the most generous, the third dimension is just irrelevant decoration. More commonly, it is, as Maarten clearly explains, unjustifiable distortion.
Even those broadly familiar with fairly recent critiques of pie charts and their 3D cousins may not know that such critiques go back a long way.
Here are some quotations from Willard C[ope] Brinton. 1914. Graphic methods for presenting facts. New York: Engineering Magazine Company. (Note that the term "pie chart" lay a little in the future when he wrote.)
"The circle with sectors is not a desirable form of presentation" (p.6)
-- as not so flexible as divided horizontal bar
"The more easy reading of the wedge or sector chart is ... largely due
to habit" (p.6)
example criticism of dimensional ambiguity:
"it is impossible for the reader to tell whether the diagram is drawn on
the basis of one dimension, two dimensions, or three dimensions ....
Methods like this cannot be too severely condemned." (p.21)
"Graphic comparisons, wherever possible, should be made in one dimension
only." (p.22) (e.g. the use of bars of different lengths)
another criticism of dimensional ambiguity:
"it would be impossible for the average reader to tell whether this
chart was drawn on the basis of height or on the basis of area .... This
chart is a typical example of thousands of illustrations used by the
popular magazines and even by some of the more pretentious reference
seasonal variation clearer with rectangular than circular plots (p.80):
"This type of chart should be banished to the scrap heap. Charts on
rectangular ruling are easier to draw and easier to understand"
"Avoid using areas or volumes when representing quantities.
Presentations read from one dimension only are the least likely to be
Maarten Buis and Grace Jessie (2)
> thank you for your explanation, though I do not quite
> understand, especially "...distort the angles...".
> The graph is very common, isn't it?
Yes, it is common (in certain areas) and it is also very wrong.
Look at a pie chart: how does it convey information? Large proportions
get slices with large angles at the point at the center of the pie, and
small proportions get slices with small angles at the point at the center
of the pie. Now think about what happens when we create the
illusion of 3D. We do that by distorting angles. Think of how we display
a cube, we know all the angles are 90 degrees, but when we draw a cube,
they certainly are not 90 degrees when creating the illusion of 3D. This
distortion of angles is necessary when we want to create the illusion of
3D, and that is why a 3D pie chart is such a bad graph: it distorts the
very information we are trying to convey.
> By the way, whether Stata can only draw plane graphs?
There some possibilities of doing things in 3D in Stata, but not many.
Maarten Buis and Grace Jessie (1)
> Could anyone tell me how to graph pie 3d-solid in Stata, as the website
The information contained in this graph is encoded in the angles between the slices. By giving the illusion of 3D you need to distort the angles, thus distorting the very information you are trying to display. So it is a very good thing that it is not possible to create such a graph in Stata.
* For searches and help try: