Abstract: The present invention is embodied in a system and method for automatically detecting desired clusters within a data space with a radial spanning device. The radial spanning system and method finds, within a data space, connected clusters or “blobs” defined by a cluster of data points. Given a probability density on data points for class inclusion and a seed point in the data space, the radial spanning system and method of the present invention finds an approximately, strongly connected cluster of data points of that class. Exploratory spokes are traversed radially outward from the seed point. Each spoke has a start point (seed point) and a final endpoint and is governed by forces based on underlying data space data points, class probability densities, internal expansion, and interspoke springs. The final endpoints of the spokes are connected to form a polygon that circumscribes the cluster of data points found.