[R] 2d plot with modification of plotting symbol to indicate third dimension.

[Prof Brian Ripley]

On Thu, 13 Jan 2011, Bert Gunter wrote:

> Duncan:
>
> I must humbly disagree. Here's the problem: in order to accurately
> represent the value, the "point" = circle _area_ must be proportional
> to the value. That is, the eye "sees" the areas, not the radii, as the
> point "size." A delightful reference on this is Howard Wainer's 1982
> or so (can't remember exactly) article in THE AMERICAN STATISTICIAN,
> "How to Graph Data Badly" (or maybe "Plot" Data).

('Display' Data, 1984).

In a pseudo-3D plot, such as scatter3dplot, size perception can be 
even more complex.

> Anyway, using cex, I have no idea whether a point drawn with cex =
> 1.23 is 1.23 times the area or radius -- or neither -- of a point
> drawn with cex =1. Indeed, it might vary depending on the
> implementation/OS/graphics fonts.

It will be approximately 1.23 times the radius.  For the default pch=1 
(and we seem to be talking about circles), in the graphics devices 
points() and symbols() will call the same graphics primitives for 
circles.  Now circles will have their sizes rounded to something, 
maybe the nearest pixel, maybe 0.01bp, maybe something else, so you 
may not have much control over very small circles, but you will over 
those of diameter about 5mm or more.

> So it seems better to me to "draw" the point with symbols(), where 
> you can have complete control over the size.

In some ways this goes back to S, where AFAIR (and ?points suggest) 
pch=1 was an octagon and symbols() was the only way to draw circles.

> Obviously, let me know if I'm wrong about this.
>
> Cheers,
> Bert
>
> On Wed, Jan 12, 2011 at 9:11 PM, Duncan Murdoch
> <murdoch.duncan at gmail.com> wrote:
>> On 11-01-12 9:33 PM, John Sorkin wrote:
>>> I would like to plot 3-dimensional data on a two-dimensional scatter-plot.
>>> Is there a way I can automatically modify the plot symbol (e.g. changing
>>> size or color) to indicate the value of a third variable? E.g. How can I
>>> plot weight vs. age and indicate the value of muscle mass for each value
>>> weight-age pair by making the plot point proportional to the subject's
>>> muscle mass?
>>
>> Just set cex to the other variable.  For example,
>>
>> plot(1:10, 1:10, cex=1:10)
>>
>> Similarly, col and pch can be set to vector values, which are recycled
>> through the points.
>>
>> Duncan Murdoch

> -- 
> Bert Gunter
> Genentech Nonclinical Biostatistics
> 467-7374
> http://devo.gene.com/groups/devo/depts/ncb/home.shtml

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595