During last years, Information Viuslization became an active field of research. As the quantity of information stored in databases increases, visualization systems becomes more and more necessary. In my research, I focus on the visualization of relationnal data. To visualize such data, one usually represents them as graphs.
To produce efficient visualization techniques (or systems), I focussed during my phd thesis on three axis: graph decomposition, graph drawing and experimental evaluation. First axis consists in clustering the graph into groups of elements sharing one (or several) properties. Main advantage of the decomposition is that it allows to build a visual abstraction of the graph and thus to reduce the number of displayed elements. Second axis consists in associating coordinates to the vertices and the edges of the graph in order to display them in a readable and meaningful manner. Finally, last axis is the evaluation of the quality of the visualization techniques (or systems) by an experimental evaluation.
One of the new aspect of my research is about interaction and more precisely about exploration of large graphs. In case of large data, it is necesserary to find a good trade off between the number of displayed elements (and therefore the level of abstraction), and the rendering speed (to support the exploration process).