[R] factorial map?

[Sacha Viquerat]

hello! i'm dealing with the following: i've collected a factor 
covariable at irregularly placed sampling points along a line with 
spatial informations, i.e.: dataset<-c(x-coordinates, y-coordinates, 
level-of-factor)
the factor describes the density of vegetation between 0 (no ground 
cover) and 5 (almost complete cover). id like to produce a map similar 
to the ones the akima package enables me to (function in akima is 
called: interp(x,y,z-value...), then use image(...) on the resulting data).
but: as the akima library works suitable with continuous z-values 
(topography, air pressure, whatever is measured can be described a 
number), my resulting map looks terribly patchy (of course, since the 
levels of the factor cannot be interpolated) and contains no valuable 
information. thus my question: i imagine something like a "growing-cells 
mechanism", which lets every sample grow until its "borders" touch the 
borders of an adjacent sample, resulting in something like a 
honeycomb-structure. that would give me a cellular pattern of my factor 
spread across the map!