When dealing with statistics, the description of complex reality often faces the difficulty of describing multidimensional phenomena through punctual indicators and quantitative data. Moreover, uncertainty is inherent in the definition and manipulation of many of the concepts, which can neither be ignored nor neglected during analysis. In order to observe and explore uncertain and multidimensional components, it is first necessary to create concrete images, thus encoding these components visually. The challenge then becomes how to give shape to these representations – these intuitions – that traditional tools of statistical visualization, primarily focused on punctual data, are unable to show. Selecting poverty as the field of investigation, this study focuses on a fuzzy multidimensional approach that is able to deal with the uncertainty of the phenomenon. The study proposes tools for the visualisation and investigation of the data structure and latent relations between different dimensions, while retaining inherent instability.