I am looking for an algorithmic way of handling this situation, as I can do it manually with the ruler and some time, but it would be nice to find a programmatic to have Arc do the work for me.
I have a bunch of point features, and these point features can be grouped into a new single feature, but only if A) the individual features are <20m apart and B) there are >4 individual features within that 20m of each other (not just 20m from a centroid).
I can buffer the individual features 20m to see which neighbours they touch, but then I need to differentiate those overlaps into 20m cluster groups.
What I don't want is a situation where PointA, B, C, D, E are all 20m from each other and should be grouped, but PointF is <20m from point E but >50m from PointA and there are no other points around, so I need a method which can understand when an individual point should be grouped because it is <20m from another point, but then figure out if there are enough other points around to make a group (>4> and then decide which group it should belong to, if there are two grouped clusters it could fit into, based on which centroid it is closest to.
Does that make sense?
I'm using ArcGIS Desktop 10.0.