Abstract
Visualization of information is a promising approach to the exploration, analysis and understanding of data. For the purpose of this thesis, the various collaborations with the end user of our method ( Biologists and Geographers) have enabled us to elaborate new visualizations and interactions adapted to their specic needs. The notion of "focus+context" exploration is central in our case. We have approached the analysis through both the visualization of multidimensional data and hierarchical graphs. A method which allows one to visualize and interact with multidimentional data is proposed. This method enables visual identication of nonlinear correlations through the help of a matrix of graph correlation. We hereby present a technique which allows the visualization in an interactive manner of hierarchical structures in the form of directed acyclic graphs (DAGs). The results of the formal valuation of the method are presented. An extension to more massive date through the use of "focus+context" visualization is thus proposed.