A method based on small area data analysis was developed to build a health risk classification for the Greater Rio de Janeiro Metropolitan Area. The approach uses 1991 census data and studies data pertaining to sanitation, ownership and type of housing, size and occupancy of the household, demography, schooling, and income. Principal component analysis applied over each dimension allowed for the choice of 15 variables, which summarized most of the observed variances. Additional analysis with these variables suggested that just six variables are sufficient for the construction of a classification using k-means method of multivariate cluster analysis. Five classes were obtained: (A) high income; (B) lower income; (C) poor; (D) low schooling and income; (E) low-level access to sanitation. The existing inequality in each of the geopolitical established areas was clearly identified. The proposed method allowed for the construction of compound indices to evaluate quality of life, based on widespread and easily obtained data (the census). Moreover, the method contributed to the detection of socioeconomic inequality, identifying, not only the larger poor regions but also the small excluded areas.