Real social networks can be described using two dimensions: first a structural dimension that contains the social graph, e.g. the actors and the relationships between them, and second a compositional dimension containing the actors'attributes, e.g. their profile, their textual interactions, etc. Each of these dimensions can be used to explain different phenomena occurring on the social network, whether from a connectivity or an individual perspective. In the case of community detection problem, exploiting only the structural dimension discards the compositional data and valuable information is not used. In this paper we claim that the integration of both dimensions would allow researchers to analyze the network. We present a novel approach to the community detection problem using the integration of the two dimensions composing a social network. A contingency matrix is generated from the agreement between two partitions, one extracted from the structure and one from the composition variable. The resulting partition shows interesting properties, such as denser and more homogeneous groups of actors, revealing new types of communities to the analyst.