Looking to spice up my statistics home work. I've figured out how to chart a
normal distribution curve using a scatter plot. How can I pick a point or
two points on the curve and shade between them to show an area?

Robb -

I think the easiest way is to use error bars (appropriately formatted).

I've posted an example, ShadedNormalDensity.xls, on
http://www.mikemiddleton.com in the bottom left corner of the home page.

As I recall, another way is a combination chart using an XY (Scatter) series
and an Area series.

- Mike

"Robb" wrote in message
...
Looking to spice up my statistics home work. I've figured out how to
chart a
normal distribution curve using a scatter plot. How can I pick a point or
two points on the curve and shade between them to show an area?

On Sun, 11 Feb 2007, in microsoft.public.excel.charting,
Robb said:

Looking to spice up my statistics home work. I've figured out how to chart a
normal distribution curve using a scatter plot. How can I pick a point or
two points on the curve and shade between them to show an area?

Make a combination chart. One of the series to be your regular curve as
a scatter chart, and the other to be an area chart. You'll have to
juggle the scales a bit to get them to match up in space.

Alternatively, you could go for an all Area chart solution instead,
using the Time-scale Category (X) axis trick, which is basically a way
of exploiting the fact that Microsoft first failed to implement an
interval scale X axis for most of its charts, then responded to business
demand for a time interval X axis by providing an interval scale *only*
for time.

This lets you treat an Area chart, normally a strictly category axis
style, as an XY chart. You'll need to double up the rows where you
change over from shaded to unshaded, because the line has to go
vertical, from zero up to the point it's at when you start shading, then
go vertically down back to zero when you stop shading.

I'll leave the details for you to work out from here. You'll also have
to make a fake X axis, because the dates on the time scale really don't
work for a normal distribution (specifically, dates can't be negative).

--
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.