Testing a Methodologic Approach to Territory Categorization Using Census Data

Abstract

The more traditional perspective in Geography views territory as a dichotomous reality of sociodemographic and biophysical characteristics yet, literature states that both territorial dimensions coexist and interact. The number of indicators for each territorial dimension is vast which makes the territory categorization a complex task. Despite it, territory categorization generates knowledge about the differences between territories that improve the planning process and make it more directed and strategic. This work main goal is to test a methodologic approach to categorize Lisbon Municipality area at the statistical section level, integrating information on the physical and social aspects. For this purpose, multivariate and complementary statistical analysis techniques will be used: Principal Component Analysis (applied to the physical and social aspects of the territory) and Hierarchical Classification Analysis (based in the factors extracted in the Principal Component Analysis). The resulting clusters will then be mapped and its distribution will be tested for spatial autocorrelation.