Companies in dynamic environments have to react to certain market events. Reactions can be short-term and influence the behavior of running process instances or they can be mid-term or long-term and cause the redesign of the process. In both situations, insights into the process flow are necessary and provided by Process Data Warehouse Systems. This paper proposes to derive the data warehouse structures from the meta model of the BPMN (Business Process Model and Notation), the actual de-facto standard of workflow languages. The resulting data structure is generic in order to be portable between application domains and to be stable in case of changing workflows.