[R] Spatstat - coordinates in observation window

Rolf Turner rolf.turner at xtra.co.nz
Fri Mar 16 20:31:46 CET 2012


On 16/03/12 23:02, Lucie V wrote:
> Dear R users,
>
> I wish to run spatial point pattern analysis (e.g. pair correlation function, mark correlation function) for which I need to create an observation window (window=owin) from which the spatial analysis is generated. The command I used to create this observation window as follows:
>
> X1<- ppp(x, y, window=owin(c(80.58,144.96),c(101.06,165.13)), unitname=c("metres","metres"), marks=dbh)
     Surely this should read

         unitname=c("metre","metres")

     i.e. singular then plural.
>
> I managed to create the observation window with this command. However, my data (x,y coordinates of objects in a square plot) didn't fit into this observation window's frame. It seems like I need to 'move or twist' my data by a certain angle so that they fit within the border of the observation window's frame. I'm not sure how to do that. Is there a command of function for that? I would be very grateful for any help or suggestions.
>
> I hope I described my problem so that it is understandable.  Thank you, L.

Why did you choose to use the window owin(c(80.58,144.96),c(101.06,165.13))
if that is not the window containing your points?

In any analysis of spatial point patterns a window should be specified 
(a priori)
and this window should consist of the region in which points were 
observed or
looked for.

Remember that there is information in *where the points aren't* as well 
as in
where they *are*.  You can't tell where points aren't unless you know 
where they've
been looked for.

So your analysis would appear to be on a shaky foundation from the start.

That being said, it is very easy (though unsound and misleading) to 
construct
a window containing all your points in an a posteriori fashion.  Tools 
to do this
include the functions ripras() and clickpoly().  Note that a window need 
*not*
necessarily be a rectangle!

There are also tools for adjusting and transforming existing windows.  See
affine(), rotate(), shift(), erosion(), dilation(), and expand.owin().

Transforming the observed points so that they fit into a given window makes
no sense at all, as far as I can discern.

     cheers,

         Rolf Turner



More information about the R-help mailing list