Handling workflow transactional behavior remains a main problem to ensure a correct and reliable execution. It is obvious that the discovery, and the explanation of this behavior, would enable to better understand and control workflow recovery. Unfortunately, previous workflow mining works have concentrated their efforts on control flow aspects. Although powerful, these proposals are found lacking in functionalities and performance when used to discover workflow transactional behavior. In this paper, we describe mining techniques, which are able to discover a workflow model, and to improve its transactional behavior from event based logs. First, we propose an algorithm to discover workflow patterns. Then, we propose techniques to discover activities transactional dependencies that allow us to mine workflow recovery techniques. Finally, based on this mining step, we use a set of rules to improve workflow design.