Given a set of values (x, y, z) = (x-coord (long), y-coord (lat), revenue), is there an algorithm similar to k-means clustering that will help find the most revenue concentrated clusterings of a fixed radius r and allowing the number of clusters to vary?
I understand that what I am laying out as the structure of this model is somewhat ambiguous and not entirely concrete, so slight alterations are welcome, but am essentially looking to find the must revenue dense areas.
You can try to compute kernel density plots for spatial point patterns, see this link: density.ppp function in the R spatstat package. Your $r$ would be the radius of whatever kernel you choose and let the intensity of the kernel be dependent upon revenue. The output of this process is a surface may analyze further, it will give you some idea of where revenue is concentrated, but it will be quite dependent upon your sampling method.
Edit: You may also want to try to compute different types surfaces to gauge revenue. Inverse distance weighting, for example. Spatial interpolation like Kriging might also be useful. A small list of different methods is here;