I just wanted to know if its possible to perform linear transformation of a directional data (circular/ spherical) e.g- wind direction, earths magnetism, etc. It would be great if anyone can share relevant links. And also is it possible to implement it in r?


Regards, Lesnar

  • $\begingroup$ What do you mean by 'linear transformation'? $\endgroup$ Mar 17 '16 at 8:33
  • $\begingroup$ @Kees Mulder By 'linear transformation', I mean that is it possible to make a directional data linear or euclidean? $\endgroup$
    – Lesnar
    Mar 19 '16 at 5:17

For data on the circle, simply take x and y as the cosine and sine of your angle, respectively.

For higher dimensions, see Wikipedia




You can simply plot a circular variable as a linear plot. For that, you don't need to do any transformation! For example -

library(circular); View(wind)

You can plot the original data without telling R that it is a circular data.

plot(wind, pch=16)

Now you are telling R that it is a circular data class and these are the specifications. Like - these data points are basically angles, measured in radians and in clockwise direction.

windc <- circular(wind, type="angles",units="radians", zero = pi/2, rotation = "clock")

Now plot. It will generate a circular plot.

plot(windc, cex=1.5, bin=720, stack=TRUE, sep=0.030, shrink=1)

For more information see Circular statistics with R book.

  • $\begingroup$ +1 The question was very unclear. This is at least helpful. $\endgroup$ Jan 16 '17 at 13:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.