FANOVA graphs have a broad range of application in model interpretation, sensitivity analysis and analysis of computer experiments. They give a clear and easy to understand visualization of function structures, reveal block-additive decompositions and also effectively improve Kriging model predictions by kernel adaption. These several fields are combined in a new released R package fanovaGraph. This tutorial presents the implemented methods step by step together with theoretical background and illustrations. Additionally it includes methods in simulation and the crucial question of threshold decision, where graphical solutions are suggested.
