Social networks have been studied over the years in different areas of knowledge in order to understand various phenomena. Collaboration networks are social networks in which relationships represent some kind of professional collaboration among people. The study of collaboration networks can help identify members or groups that are important and influential within that community. Intuitively, relationships in collaboration networks have different intensities that can be exploited to better characterize phenomenon. This work is divided into two parts. The first part is a study of the topological properties of two collaboration networks, the global collaboration network and the Brazilian collaboration network of authors of scientific papers within the area of Computer Science. Among the properties studied, we focus on the characterization of the intensities of relationships in these networks . The second part presents a ranking metric for vertices and groups of vertices based on the intensities of their relationships. Using the proposed metric and other more classical metrics, we rank the postgraduate Brazilian programs and researchers in Brazil in Computer Science. The evaluation of the proposed metric was performed by comparison with subjective evaluations of researchers and programs made by CAPES and CNPq. The results show the effectiveness of the proposed metric in identifying influential members and groups when compared to another metrics in the literature.