Analysis of social vulnerability to natural risk using multivariate statistical techniques

Abstract

The combination of a human population and a potentially destructive agent does not necessarily produce a disaster. For disaster to become inevitable the population must be vulnerable. This study analyses social vulnerability to natural risks using multivariate statistical techniques, specifically, principal component analysis and k-means cluster analysis. The combination of these techniques avoids the need for a priori weighting of variables. This methodology has been applied as a case study in the city of Almería using a set of variables that influences the vulnerability of the population. Social vulnerability has been conceived holistically, encompassing its different dimensions, and avoiding an excessive simplification of the complexity of social reality. Therefore, it has been considered that there are multiple social, economic, and cultural factors that influence the appearance of a disaster. The results obtained have made it possible to identify the areas of greatest social vulnerability, highlighting that around 32 percent of the population is highly or very highly vulnerable. Finally, the proposed methodology has proven to be adequate for the study of social vulnerability and is also replicable in other areas of study and at different scales.