Ordinary Kriging applied to spatial analysis of intra-urban inhabitability

Intra-urban inhabitability is a wide concept which is related to not only the dwelling quality per si, but also to the quality of its surrounding. When an intra-urban inhabitability analysis explicitly considers the spatial arrangement of the avalailable data, it is possible to obtain more significant results than those obtained with techniques which consider the observational units as independent samples. Ordinary Kriging is a geostatistic interpolation method which is considered as the best linear unbiased estimator due to its capability to access the incertatinty degree in the estimation of non-sampled values, to minimize the error variance, and to permit the identification of the more appropriated neighbourhood and weights for the inferential procedure. This paper aims to access the results obtained from the application of the Ordinary Kriging method in the spatial analysis of intra-urban inhabitability of the city of Recife (Brazil). Firstly, it was created inhabitability indexes from Demographic Census 2000 data. Then, it was done an exploratory analysis of these indexes. Finally, it was definied an anisotropic semivariogram model to measure and map the inhabilitability spatial autocorrelation in Recife. The results reveal the principal spatial trends and clusters in the city.

Surface generated through interpolation by Ordinary Kriging of an Inhabitability Index (IH) created from Census 2000 data of Recife, Brazil . The darker the green areas, the higher the IH values. The darker the red areas, the lower the IH values.