Re: operations on (X,Y) series

[Jean Bréfort <jean brefort normalesup org>]

Le vendredi 04 novembre 2005 à 16:43 +0100, Frederic Parrenin a écrit :
> Hi,
>
> > > > > I would like to do various operations on (x_i,y_i) series in gnumeric,
> > > > > e.g.:
> > > > > - resampling, by simple interpolation or by integration, with a linear,
> > > > > spline or staircase interpolant
> > > > > - filtring by various methods
> > > > > - smoothing by various methods
> > > > > - fitting
> > > > > - principal components
> > > > > - Periodogram
> > > > > - etc.
> > >
> > > > Do you want to implement that in sheets, in graphs, or both?
> > > > Both sheets and graphs have already more regressions than linear.
> > >
> > > I thought in sheets.
> >
> > OK, but I feel many of them would be nice in graphs too.
>
> That is true, I did not think about that.
> But at least, when they are in sheets, you can plot the result on
> graphs.

Of course ;-)

> > > There is one specific thing I would like to do in graphs.
> > > It is a visual correlation tool for two XY series. With that tool, you
> > > can stretch the X-scale of a serie to fit with another serie.
> > > This tool is available in a free software called AnalySeries,
> > > unfortunately only available on the Mac.
> >
> > This might be possible with two x-axis. It might be possible to adjust
> > one of the axis scale (I mean automatically adjust, not by hand, of
> > course).
>
> I am not sure we speak about the same thing.
> I meant, strecting the X-axis by e.g. a linear by part transformation.

This is a bit more complicated, but not impossible.

> > > This software also do the various math operations on XY series I
> > > described above. Its author (Didier Paillard) is a colleague/friend of
> > > me, so we could collaborate with him.
> > > For example, we could extract the math from AnalySeries into a library
> > > and use it in this gnumeric plugin.
> >
> > The code must be release under the GPL.
>
> It is under the CeCill licence. It is compatible with the GNU-GPL.
> More info here:
> http://www.cecill.info/

Great.

Cheers,
Jean