Data in BIM projects include information on material characteristics, dimensions, energy simulations, equipment maintenance, occupancy patterns, space functions, etc. In these projects, the process of creating and retrieving building data is crucial for the larger construction project. The detailed review of ongoing BIM projects in France revealed how much the aforementioned process had gained weight, to the point of becoming a complete project in itself, the "building data project". We formulate the hypothesis that building data and their classification are a suitable observatory of the system of actors working on the building construction process and their interests. Through the building data project, a whole set of new concerns are integrated into the modelling process, which was previously mainly focused on formal and functional aspects of the building. We therefore considered data classification and selection as an opportunity to explore emerging interests and requirements in the building design process. Following the analysis of databases related to 12 BIM projects in the design phase, we propose a tool for the classification of building data into multiple categories which reflect the actors’ main concerns and interests.