Abstract: The explosive growth of academic publication data makes it very difficult to study research groups.How to evaluate and compare the competitiveness of research teams,and identify core groups is becoming more and more important.By leveraging the publication data,this paper proposes a hierarchical index system for depicting the competitiveness of research groups.We introduce an enhanced clique percolation method(CPM)by using the clique-strength to find research groups in co-author network.We design and implement a visual comparison approach to support interactive exploration of the competitiveness of different research groups in a field.Experiments and expert reviews on the case in the field of astronomy demonstrate the usefulness and effectiveness of our approach.