Visualization for preserving internal structures in scale-free social networks based on graph databases

Visualization of social networks intuitively reveals various structural characteristics. Exposing internal structures helps further uncover the complex and intrinsic characteristics of complex networks during the visualization process. In this paper, we propose a multi-level visualization approach which preserves internal structures. First, we design a dense group detection approach through k-core decomposition and the iterative removal of maximal k-core to detect dense groups. Then, we introduce a community detection method based on these dense groups to uncover community structures. Finally, we propose a hierarchical community identification method to handle large communities. The entire process is implemented using a graph database, enabling repeated visualization. Experiments demonstrate that our graph database implementation of k-core decomposition and community detection algorithms achieves superior computational efficiency. Furthermore, experimental results indicate that the proposed parameters, community detection algorithm, and hierarchical community detection algorithm demonstrate higher effectiveness in the VFM (Visual-Friendly Modularity) metric across different datasets. Furthermore, we investigate the non-trivial structures in real-world social networks based on the proposed framework. Statistical properties observed in these networks allow us to derive conclusions about the corresponding patterns.