[R] Labeling barplot bars by multiple factors

Jim Lemon jim at bitwrit.com.au
Fri May 29 12:40:40 CEST 2009


Thomas Levine wrote:
> Ah, that makes sense. But now another two issues have arisen.
>
> Firstly, the error bars look like confidence intervals, and I'm pretty
> sure that they are but does some document verify this? I suppose I
> could check the code too.
>
> Secondly, I just read about how dynamite plots should be avoided. It's
> quite easy to turn the dynamite plots into dot plots with Inkscape,
> but is there an equivalent function that generates _hierarchical_ dot
> plots?
>   
Hi Tom,
The "error bars" in hierobarp are whatever you specify they are. In the 
examples, both standard error and quantiles are illustrated. If you 
really don't want bars, it might be possible to display hierarchical 
dotplots, but I doubt that it would visually convey the hierarchical 
breakdown as well as the overplotted bars.

Jim




More information about the R-help mailing list