On Tue, 18 Sep 2007, in microsoft.public.excel.charting,
itkovian said:
I have a set of data I wish to graph in a scatterplot. Each data point
has two identification codes, coming from two separate groups, e.g., L
= {a,b,c,d,e} and C = {1,2,3,4}. So you might have points

I wish to use both colour and shape to identify the data points, e.g.
all a's should be yellow, all b's red, etc. and all 1's should be a
triangle, all 2's a square etc.

It seems to me that you are coding the codes. Why would you create a
code 1, 2, 3, etc., and then represent them by triangle, square, etc.?

If you use one of the available add-ins for custom labels, you could use
the labels 1,2,3 directly on your graph without requiring the viewer to
perform an act of translation. Position the labels centre,centre and
format the scatter series to have no marker. Then you can represent a,
b, c, by the font colours of the labels, having arranged your data so
that each a, each b, and so on represent a distinct range.

Using custom labels can be a very powerful intuitive graphing technique,
for instance when presenting a scatter graph of two variables across
fifty US states, using the two-capital-letter abbreviations of the
states as the markers. Here's an example using TV stations in two
different colours:

http://junkcharts.typepad.com/junk_c...nd_median.html

--
Del Cotter
NB Personal replies to this post will send email to ,
which goes to a spam folder-- please send your email to del3 instead.