| Re: scatter plots make large PostScript files [message #64502] |
Fri, 02 January 2009 10:47 |
pyoachim
Messages: 2
Registered: January 2009
|
Junior Member |
|
|
On Jan 1, 1:35 pm, Ed Hyer <ejh...@gmail.com> wrote:
> Too large, in fact. In this case, I want the scatter as an underlay to
> a line plot showing the averaged data. But the resulting EPS file is
> just too huge. Any tips/tricks/brilliant ideas on how to get the plot
> I want with a manageable EPS file size?
What I often advocate in these situations is using hist2d and contour
to plot the underlying data as a contour plot.
|
|
|
|
| Re: scatter plots make large PostScript files [message #64510 is a reply to message #64502] |
Fri, 02 January 2009 05:58  |
Jeremy Bailin
Messages: 618
Registered: April 2008
|
Senior Member |
|
|
On Jan 2, 3:54 am, anand <jana...@gmail.com> wrote:
> On Jan 2, 12:35 am, Ed Hyer <ejh...@gmail.com> wrote:
>
>> Too large, in fact. In this case, I want the scatter as an underlay to
>> a line plot showing the averaged data. But the resulting EPS file is
>> just too huge. Any tips/tricks/brilliant ideas on how to get the plot
>> I want with a manageable EPS file size?
>
> you can use the keyword NSUM while plotting. it is like binning the
> available
> data. NSUM=10 would reduce the number of data points in the scatter
> from
> say, N, to N/10.
>
> Anand.
Note that because NSUM averages, the resulting plot does not represent
the original scatter (and it goes without saying that outliers
disappear). I find it's often better to randomly sample the original
points. If there are noriginal points in the data set and you want to
plot nplot of them:
randomsort=sort(randomu(seed,noriginal))
plot, x[randomsort[0:nplot-1]], y[randomsort[0:nplot-1]], psym=3
-Jeremy.
|
|
|
|
| Re: scatter plots make large PostScript files [message #64512 is a reply to message #64510] |
Fri, 02 January 2009 00:54 |
anand
Messages: 11
Registered: December 2008
|
Junior Member |
|
|
On Jan 2, 12:35 am, Ed Hyer <ejh...@gmail.com> wrote:
> Too large, in fact. In this case, I want the scatter as an underlay to
> a line plot showing the averaged data. But the resulting EPS file is
> just too huge. Any tips/tricks/brilliant ideas on how to get the plot
> I want with a manageable EPS file size?
you can use the keyword NSUM while plotting. it is like binning the
available
data. NSUM=10 would reduce the number of data points in the scatter
from
say, N, to N/10.
Anand.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
