[R] cannot use package RcmdrPlugin.plotByGroup

Fox, John jfox at mcmaster.ca
Sun Jul 23 14:41:16 CEST 2017


Dear Lu Wei,

> -----Original Message-----
> From: R-help [mailto:r-help-bounces at r-project.org] On Behalf Of Lu Wei
> Sent: Sunday, July 23, 2017 12:27 AM
> To: r-help at stat.math.ethz.ch
> Subject: Re: [R] cannot use package RcmdrPlugin.plotByGroup
> 
> On 2017-7-23 6:21, Fox, John wrote:
> > Dear Lu Wei,
> >
> > I'm the author of the Rcmdr package but not of
> > RcmdrPlugin.plotByGroup, so if there really is a problem -- which
> > isn't apparent from your message -- then you may wish to contact the
> > package author.
> >
> > First, as a general matter, menu items in the Rcmdr are grayed out if
> > they're inappropriate in the current context. Rcmdr plug-in authors
> > can use the same mechanism. So, for example, most menu items are
> > grayed out before you read a data set, and menu items that require a
> > categorical variable will be grayed out if there are no factors (or
> > character variables or logical variables) in the active data set.
> > ...
> 
> Thanks for the reply. I change to another data set and the menu is available. I
> did not know the type of variable -- I am a newbie who is searching a quick
> solution, that's why I try R-commander. 

The Rcmdr comes with an introductory manual, accessible via "Help > Introduction to the R Commander". This and a lot of other information is in the manual, and you might try reading it.

> But after some test I find that plugin's
> output is not exactly what I want.
> 
> I am seeking a way to visualize data by plotting grouped box and distribution
> chart. The latter may be called Q-Q chart, which I just learned from R, though it
> seems not exactly the same as I have seen below. I hope the groups are in the
> same chart for comparison, like this:
> http://cubeupload.com/im/WFc4CC.png
> 
> or even plot a set of variables:
> http://cubeupload.com/im/wK69NN.jpg
> 
> These charts are produced by Minitab. The norm-axis is not the same as Q-Q
> chart in R, but I think they are equivalent. Is it right?

I'm afraid that I can't say because I don't read Chinese. It looks like quantiles of some sort are on the vertical axis, which is labelled as percentiles rather than with values of the variable; if these are quantiles of the standard normal distribution, then the graphs are normal QQ plots. In the Rcmdr, normal quantiles appear on the horizontal axis, which is more common, and the scale is labelled with values of the variable.

> 
> I hope you or the author of RcmdrPlugin.plotByGroup could implement the
> feature in the picture I posted. This "distribution chart" or Q-Q chart is the most
> I want from a statistic tool, since it gives full information of experiment data,
> and putting groups together in a chart for comparison is the ultimate
> judgement.

The Rcmdr uses the qqPlot() function in the car package. In the next version of the car package, qqPlot() will be able to plot by groups, and when that happens, I'll incorporate the feature in the Rcmdr.

More generally, however, if you learn to write R commands you'll be able to make customized plots. It would be easy to duplicate the plots that you show, for example.

Best,
 John

> 
> Thank you for making R an easy tool to learn and use. Great job.	 
> 
> --
> Regards,
> Lu Wei
> PGP key ID: 0x A12F EF75 92CC E1EA
> 
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.



More information about the R-help mailing list