The intention of this thesis is to give a better understanding of the way that subduction zones evolve, to examine the way that different subduction parameters affect each other and finally to demonstrate how visualization can be used as a tool to provide deeper insight into such zones. The first part, describes the characteristics and the theory behind the formation and evolution of the areas that the phenomenon of subduction takes place. It can be especially useful to readers who do not have extensive previous knowledge on this subject. The second part, uses 20 measurable parameters of subduction zones to develop statistical models in order to reveal correlations and tendencies within geological observations around zones of subduction. This models are created using multiple linear regression with the help of the R statistical software environment. The third and final part, deals with the visualization of geological phenomena by using software that computer science has developed for geoscience. For this purpose, bathymetrical reconstruction of the Cascadia slab is performed and visualized, using the previously acquired models and the 4DPlates plate reconstruction software.