This tutorial has two major aims: The first one is to show the general workflow of how land cover classifications (or similar tasks) based on satellite data can be performed in R using machine learning algorithms. The second important aim is to show how to assess the area to which a spatial prediction model can be applied ("Area of applicability", AOA). This is relevant because in spatial predictive mapping, models are often applied to make predictions far beyond sampling locations (i.e. field observarions used to map a variable even on a global scale), where new locations might considerably differ in their environmental properties. However, areas in the predictor space without support of training data are problematic. The model has no knowledge about these environments and predictions for such areas have to be considered highly uncertain.