2004 年 12 巻 2 号 p. 155-164
One of the most fundamental spatial analytical methods on a two-dimensional plane is the cell count method. This paper extends the ordinary cell count method defined on a plane, called the planar cell count method, to a method defined on a network, called the network cell count method. These two methods are applied to the same distribution of points in a small study area in Shibuya, Tokyo. The statistical tests by the goodness-of-fit test and a spatial autocorrelation analysis reveal that the planar cell count method tends to overestimate the extent of dispersion in comparison with the network cell count method.