On May 7, 10:53 pm, Spon <christoph.b...@[EMAIL PROTECTED]
> wrote:
> Bernat wrote:
> > for i=0, 1000 do oplot, datax[i,*], datay[i,*]
> > Drawing all the plots is a good idea for analyse the data.
>
> Something tells me there are several people you've yet to convince of
> this ;-) (myself included). Just out of curiosity, what are you trying
> to visualise in this way? I'd be interested to know how sparsely
> scattered a dataset would need to be for this method to give you
> something other than a headache. :-)
I'm with Bernat: even with large datasets, drawing all the data as
line plots can be useful as a quick way of distinguishing sparsely
populated regions of the parameter space from densely populated
regions, and for detecting outliers.
I recall that back in 1994 or so I used IDL to plot several time
series of air pollution data, with ~ 10^6 points per time series, and
saw stuff in there that others had not seen, because they were stuck
with puny spreadsheet-based plotting tools. ... That would have been
Direct Graphics, BTW :-)
|
