[rkward-devel] pie chart

[Stefan Roediger]

Am Mittwoch, 8. August 2007 09:28:17 schrieb I. Soumpasis:
> 2007/8/7, Stefan Roediger <stefan_roediger at gmx.de>:
> > Hi list,
> >
> > I added an input field for names in the pie chart plug-in. I can't
> > remember
> > the original author. Anyway, please check if you like the approach. From
> > now
> > on the user is forced to enter names (may it be the original data (->
> > numbers) or strings) in contrast to the previous way where the "name" was
> > chosen "automatically".
>
> Hi Stefan,
>
> I tried the plugin tabulating factor data but I cannot produce a pie chart.
> The problem seems to be the names (x) <- "data". Someway the users should
> be warned I think that the dimension of names (x) must be equal with the
> data they choose for names,

I only had scenarios in mind where the dimension of of names is equal to the 
data.

> or they must not be forced, just to have a 
> choice if they want to use different names.

True. I didn't consider this. This making it optional is certainly the way to 
go. Thanks for pointing that out.

>
> BTW, who is active right now?
>
>
> I am here and there. I am in a transitional fase and when I find some hours
> free I commit as far as I can.
>
> Kind Regards,
> Ilias
>
> BTW, I saw you started the lm plugin. I have some ideas about the lm - glm
> plugins and some code written in R (I have already sent a e-mail on the
> list before but did not have the time for more). If you want I could send
> you some code or we could develop together these plugins because I believe
> that they are important and the user should get the maximum out of them.
>

Sounds good. I was thinking about the "maximum" and how to achieve it. But I 
do not see it right now (what is needed ...?). If you have some input ... .

> Two points on the plugin already at the svn repository:
> 1. The Independent Variables could be more than one so if the user chooses
> one he will do lm, if he chooses more than one he could do multiple lm.

Yes of course that's basically possible but I would like to find a way to do  
some sophisticated equations like:
fit <- lm(y ~ x + I(x^2) + I(x^3) + I(x^4) + I(x^5))
But currently I don't see it from the code GUI side.

> 2. The user should get a choice of plotting. The plot (lm(x~y)) gives four
> plots. If we use par(mfrow=c(2,2)) and then the plot(lm(x~y)) could give
> use the four graphs in one plot.

I played around with that before and must say it's certainly needed but I'm no 
longer sure if the par(mfrow=c(n,m)) approach is the best. Having single 
plots of plot (lm ()) is better since it's easier to read and work with later 
on. But this could be done as option of course.