Easy.  Put it into RubberSheet and Image, with AutoAxes turned on.  I've
attached a net file.  I create a vector of random 2-D positions in the unit
square, pass it into Histogram requesting 20 buckets on a side, and render
the result - a 2-D bar chart.  For an interesting alternative, add Post(...
"positions") after Histogram and before RubberSheet.

Greg

Sharon Gibson <[EMAIL PROTECTED]>@opendx.watson.ibm.com on 10/12/2001
04:14:35 PM

Please respond to [email protected]

Sent by:  [EMAIL PROTECTED]


To:   [email protected]
cc:
Subject:  Re: [opendx-users] cluster plot



I can take x and y and create 2-vector data of the form [x,y] and then send
to
Histogram which will calculate the frequency of occurence. But then how do
I
visualize that result ? It does not appear to readily feed back into the
Plot
module.

I want to simulate the scatter plot that I already have, but with a single
colorized point at each [x,y] to indicate the number of occurences (many
plotted points would overlap eachother, otherwise). Should I ditch the Plot
module ?

On Oct 12,  2:03pm, Gregory D Abram wrote:
> Subject: Re: [opendx-users] cluster plot
>
> Isn't this exactly what the Histogram module does?
>
> Greg
>
>
> Sharon Gibson <[EMAIL PROTECTED]>@opendx.watson.ibm.com on
10/12/2001
> 01:37:50 PM
>
> Please respond to [email protected]
>
> Sent by:  [EMAIL PROTECTED]
>
>
> To:   [email protected]
> cc:
> Subject:  [opendx-users] cluster plot
>
>
>
> I'm looking for a good way to represent frequency of occurence in an x
vs.
> y
> cluster plot.
>
> I've developed an array based solution to this kind of thing using other
> packages, but I have no clue how to start in dx, since I now have fields
> rather
> than arrays.
>
> Can anyone give me a directive ? I don't know whether to look for an
> entirely
> different form of visualization, or try to force this thing into the
> array-based concept that is in my head (bin the positions, and somehow
> count
> number of occurences in that bin... ugh).
>
>
> -Sharon
>
>
>
>-- End of excerpt from Gregory D Abram




Reply via email to