ABSTRACT
In contrast to Chapter 4 on defining functional regions, within which connection strengths are the strongest, Chapter 9 delineates homogeneous regions, within which areas are more similar in attributes than those beyond. The chapter begins with an illustration of the small population problem and introduces several GIS-automated regionalization methods, such as SCHC, SKATER, AZP, Max-P, and REDCAP, for constructing regions with sufficient population to mitigate the problem. A separate section discusses the mixed-level regionalization (MLR) method, which decomposes areas of large population and merges areas of small population simultaneously to derive regions with comparable population size. A case study of analyzing breast cancer rates in Louisiana is used to illustrate the implementation of these methods. By changing the desirable number of derived regions or imposing a constraint of a threshold population, the methods can generate a series of different sizes of regions. A method with such a capacity is thus termed “scale-flexible regionalization” and enables an analyst to examine a study area at different scales and detect the modifiable areal unit problem (MAUP) or the uncertain geographic context problem (UGCoP).
