[R] Attempting to sync x-axis labels with grid lines in lattice bwplot

[Deepayan Sarkar]

On 4/20/08, Mark Coletti <mcoletti at gmail.com> wrote:
> I have created a lattice bwplot that suffers from axis label
>  overplotting.  The box plot contains grid lines, and I thought that
>  only plotting the labels for corresponding vertical grid lines would
>  not only cure the overplotting but also be more aesthetically
>  pleasing.  Unfortunately hours of poring over manuals, documentations,
>  mailing list archives, and other R-related text hasn't helped.
>
>  This following is the plot:
>
>  http://www.lychnobite.org/images/bsf.png
>
>  As you can see, the x axis labels are a mess.  :(
>
>  This is the code I used to generate the plot:
>
>  bsf.plot = bwplot( fitness ~ Gen | Crossover * Rules,
>   data = bsf.df,
>   panel = function(...) {
>     panel.grid(v = -10, h = -1)
>     panel.bwplot(...)
>
>     # cribbed some lines from lattice grid function
>     limits <- current.panel.limits()
>     print( limits$xlim )
>
>     # maybe add x axis plotting here? ... and suppress it elsewhere?
>
>   },
>   subset = Gen %in% filtered.gen,    # mask for plotting only some generations
>   par.settings = list(plot.symbol = list(pch = 20)), # dots for outliers
>   pch="|",                              # want median lines not dot
>   fill="lightgray",                     # gray stands out more
>   xlab="Generations",
>   main = "Best so far" )
>
>  I suspect that I'll have to use "at" within "scale" -- but where to
>  get the grid line coordinates?  As seen above, I cut-n-pasted some of
>  the source from "panel.grid", but then realized that there seemed to
>  be no way to get that information to the "scale/at" parameter.

A factor such as Gen here would get converted to numeric values using
as.numeric() before plotting, so you can infer the 'at' values from
that. In other words, the level "0" gets plotted as 1, "10" as 2, and
so on.

-Deepayan