From Databases to Graph Visualization

The first step of any information visualization system is to enable end user to import their dataset into the system. However, non expert user are faced to the difficult task of choosing how their data should/could be transform to be used in these Infovis systems. In that paper we address the case where end users want to use dataset in tabular format. We propose a novel method for automatic graph generation from these datasets. That method consists in first building taxonomy of dimensions. Then, that taxonomy is used to provide to user a system that enables to interactively navigate into the set of possible data transformation.