||Construction project participants constitute a complex social human network composed of a heterogeneous and fragmented set of stakeholders. The disjoint group of actors that team to work on a project constitutes collective entities, social networks at different scales in time and space. The proposed social network system is a semantic resource that leverages the communication and coordination of exchanging and sharing information. It is expected that it will enable an improvement in efficiency of the interfacing of actors and information. This semantic resource helps actors to minimize human intervention for coordination and information searching and retrieval, which are activities that demand costly resources and the use of specialized labor. Floorbook analyzes the vocabulary of the annotations on the forms of representation used in construction documentation, categorizes and models communities according to the user’s role in the shared form of representation, and makes suggestions to the users to optimize their particular world view, so that the suggested annotation is more precise and personalized. The basic rational of the approach is that the position of the users in a social network impacts their use in the system, and that the content of the annotations are part of a categorization model of a specific domain. The proposed social network system works as an effort of collective intelligence that enables the sharing of the semantics of the tags that are associated with the representations. As an effort of collective intelligence, Floorbook (1) models and extracts semantics from informal communication; (2) categorizes and models communities defined by common interests; and (3) self-learns from the history of user actions in the system to enable new value-added services, such as, for example, suggesting new candidate semantic tags as a result of the analysis of the representations to optimize the particular world view of an individual user.