Fig. 4

Correlation between content and network structure. a Correlation between network efficiency and average topic similarity in text-based communities; b the relationship between topic similarities in communities and topic similarities in the maximum connected component (denoted by S_Gcc in text-based communities); c correlation between network efficiency and the average topic similarity in topology-based communities. The size of scatter represents the size of community and the color corresponds to the level of topic similarity (green, low; yellow, high)