Application of nonsupervised classification to population data

This paper describes the processing of a subset of data from the XII Population and Household Census for the Metropolitan Zane of the Mexico Valley with the purpose of establish and describe properties of population data. In order to analyze these population data a statistical classification technique known as cluster analysis (CA) is applied. The usefulness of CA in this case resides in solving the categorization problem with basic information that is often used in land planning and zonification of the urban space. With the objective of obtaining a classification of zones that reflects the interrelation of their variables, several clustering options are considered, with number of clusters going from 8 to 70, all of them created by the k-means procedure. Lastly, the results are presented along with maps.