In this paper we use publicly available socioeconomic data from the Greater São Paulo region in Brazil to implement a multi objective analysis in order to rank socio economic risk factors relevant for Food Insecurity and Hunger. The data we use are based exclusively on regional aggregates and not on individual households as in previous studies. The datasets used have differing characteristics, hence consistently combining the available data into a single dataset and checking the consistency is not straightforward. Novel methods to address this question will be presented. Next we will infer a Bayesian Network using techniques well established in bioinformatics. Due to the limited size of the available data the application of standard methods leads to inferred networks that are potentially unreliable. In order to circumvent this problem, feature selection will be implemented via an analysis of Pareto Corners. This method is statistical in nature and is valid even when there is no explicit parametric form for the objective functions, as is the case of the data we analyze.