I'm trying to graph some data in a boxplot-like style, but I want to set
the box and whisker limits myself (rather than having R calculate them for
me). I'd like the boxes to be shaded and the whiskers to be dotted lines.
My data are set up is something like this:
min.whisker max.whisker min.box max.box species
0 200 20 37 A
0 150 0 10 B
10 55 12 20 B
etc... I have a categorical variable (species) which I'd like to use to
group the boxes along the x-axis of the graph. I tried to do this by
plotting a blank graph and adding the boxes and lines individually using
the lines() function, but it was labour intensive, inelegant, and I
couldn't get the bars in each group (species) not to overlap with one
Does anyone know how to accomplish a graph like this in R with more
succinct coding and/or a custom graphing package? Thanks!
On 2012-06-29 11:09, briannamwright wrote:
> Hi Mark,
> Thanks for your help - I took a look at ?bxp and ?boxplot. I guess I'm just
> uncertain about how to specify "z" in bxp(), not sure I'm quite
> understanding how the function works.
Just generate the list; see the Value section in ?boxplot.
It seems that all you need is a list consisting of two
components: 'stats', a matrix of the relevant numbers,
and 'names'. Use your data to construct a 5-row, 3-column
matrix. Each column should have, _in the appropriate order_,
the 5 values listed on the help page for boxplot. Your
original post did not specify medians; just make up any
number that falls inside the box and use a transparent
colour for the median line in the plot.