Visual Data Mining (VDM) aims at an easier interpretation of data mining algorithm results through the use of visualization techniques. During the last decade, many techniques of information visualization have been proposed, allowing visualization of multidimensional data. Previously, ((Chi, 2000), (Herman et al., 2000)) attempted to classify VDM techniques . However, these taxonomies do not take into account some innovative techniques based on 3D visualization and virtual environments (VEs). In this paper, we propose an exhaustive survey of recent techniques for VDM. These different techniques are detailed, classified and compared according to the following criteria : graphical encoding, interaction techniques and applications. Moreover, they are presented in tables together with graphical illustrations.