Spatial data sets, generated from a wide range of physical systems can be analyzed by counting the number of objects in a set of bins. Previous work has been limited to equal-sized bins, which are inappropriate for some domains (e.g., circular). We consider a nonequal size bin configuration whereby overlapping or nonoverlapping bins cover the domain. A generalized index, defined in terms of a variance between bin counts, is developed to indicate whether or not a spatial data set, generated from exclusion or nonexclusion processes, is at the complete spatial randomness (CSR) state. Limiting values of the index are determined. Using examples, we investigate trends in the generalized index as a function of density and compare the results with those using equal size bins. The smallest bin size must be much larger than the mean size of the objects. We can determine whether a spatial data set is at the CSR state or not by comparing the values of a generalized index for different bin configurations—the values will be approximately the same if the data is at the CSR state, while the values will differ if the data set is not at the CSR state. In general, the generalized index is lower than the limiting value of the index, since objects do not have access to the entire region due to blocking by other objects. These methods are applied to two applications: (i) spatial data sets generated from a cellular automata model of cell aggregation in the enteric nervous system and (ii) a known plant data distribution.

• Received 21 March 2012

DOI:https://doi.org/10.1103/PhysRevE.85.061908