Data science and machine learning methods find increasing application in different fields of engineering. Model building is one important learning objective in the context of machine learning. In this paper, we analyse the structure of engineering students’ first model building processes and identify five characteristic structures. In this context, a method for illustrating and analysing a model building process is introduced, the "processing graph". The processing graph and the characteristic structures provide the opportunity to describe the structure of individual model building processes and to connect these with content-related aspects in future research projects.