Windrose Percent Interval Frequencies Are Non Linear! Help!

Brittain Briber <brittbriber <at> yahoo.com> writes:

? Do you two or anyone
else out there happen to know if it is possible to specify the
 width of the intervals in windrose?? Is there
a parameter that I can pass along that would do this??
 Below is a reproducible script in which you can see
the uneven interval spacing.? And thanks again for your answers.

Best
Britt

The "Details" of ?windrose in circular says
quite explicitly that it preserves the areas of "pedals" [sic],
so it's unlikely that this is adjustable without hacking the
code. (Hmmm, maybe I will write to the maintainers and ask
them to correct the spelling of "petals" throughout ...)

  On the other hand, there are a *lot* of alternative windrose/circular
histogram-plotting functions out there in CRAN-contrib-land.
library("sos"); findFn("wind rose") finds them in packages
plotrix, openair, oce, climatol, IDPmisc, ggplot2 ... and
that's not counting rose.diag in the CircStats package.

  I would take a quick look and see if any of those packages
work for you. 

  Ben Bolker

Brittain Briber<brittbriber<at>  yahoo.com>  writes:

   Do you two or anyone
else out there happen to know if it is possible to specify the
  width of the intervals in windrose?  Is there
a parameter that I can pass along that would do this?
  Below is a reproducible script in which you can see
the uneven interval spacing.  And thanks again for your answers.

Best
Britt

   The "Details" of ?windrose in circular says
quite explicitly that it preserves the areas of "pedals" [sic],
so it's unlikely that this is adjustable without hacking the
code. (Hmmm, maybe I will write to the maintainers and ask
them to correct the spelling of "petals" throughout ...)

   On the other hand, there are a *lot* of alternative windrose/circular
histogram-plotting functions out there in CRAN-contrib-land.
library("sos"); findFn("wind rose") finds them in packages
plotrix, openair, oce, climatol, IDPmisc, ggplot2 ... and
that's not counting rose.diag in the CircStats package.

   I would take a quick look and see if any of those packages
work for you.

   Ben Bolker

Brittain,

I think that Ben is correct: you would have to hack the code.
Here's a way around that, but it just seems to show up another
problem in windrose():

   set.seed(1)
   x <- circular(runif(100, 0, 360), units="degrees")
   y <-  rgamma(100, 15)
   DF <- data.frame(x, y)
   windrose(DF, fill.col='yellow', cir.ind=NULL)

## now superpose a set of circles
   symbols(matrix(0,5,2), circles=1:5/5, inches=FALSE,
           fg="red", add=TRUE)

There is obviously a problem with the aspect ratio and I
think it's in windrose(). Here's a more pronounced look
at the problem:

   windrose(DF, fill.col='yellow', cir.ind=NULL, bins=1)
   symbols(0, 0, circles=1, inches=FALSE,
           fg="red", add=TRUE)

You could also have a look at the draw.circle() function
in the plotrix package.

Peter Ehlers

______________________________________________
R-help at r-project.org mailing list
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.