Perspectives for using virtual reality to extend visual data mining in information visualization

Although virtual reality (VR) has a huge success in increasing the quality of scientific visualization applications, there is a considerable lag in the development of VR applications in the case of information visualization (InfoVis). Some researchers in InfoVis claim that 2D representations are enough for data analysis; however, in the case of multi-dimensional datasets, other researchers indicate that studying multiple dimensions simultaneously is advantageous [1], [2], [3]. The first studies with low quality stereoscopic devices showed no advantages when performing simple tasks on simple datasets [4]. However, recent experiments using higher quality devices and more complex datasets show a huge improvement in performance [5]. Still, designing an effective 3D representation remains a complex endeavour. Brath [6] provides a list of things to take into account. This paper shows how a CAVE-like2 environment can allow the study of multi-dimensional datasets while keeping multiple 2D representations available, using genome comparisons as a usecase. Parallel coordinate plots and non-planar graph representation in 3D space are also described. Some potential applications in aerospace are also listed (this is by no means a complete listing of applications; readers are encouraged to suggest new ones).