[R-sig-Geo] center of mass for variable values in a grid

[Waichler, Scott R]

> >>  If your data is in a data frame of coordinates x, y, (and maybe z),
> >> and variable W for all cells, and all your cells are the same size,
> >> then its just the weighted average of the coordinate (assuming your
> >> dataframe is the complete grid). So:
> >>
> >>  xm = sum(d$x*d$W)/sum(d$W)
> >>  ym = sum(d$y*d$W)/sum(d$W)
> >
> > You're right, it's simple for a grid where dx and dy are constant.
>  Thanks.  But if they vary?
> 
>  Treat the masses as points centred at each grid cell and then find
> the weighted mean of X and Y as before.

Sorry, I was wrong in thinking that a regular grid is needed.  The simple weighting scheme above will work for any set of points, as long as the weight represented at each point is centered there.

Scott Waichler