BLOCKS: Efficient and Stable Online Visualization of Dynamic Network Evolution

It is computationally and cognitively expensive to observe the evolution of community structure on dynamic networks in real-time, not only because the data sets tend to be complex, but also because the visual interfaces are often complicated. We introduce BLOCKS, a simple but efficient framework to abstract and visualize the evolution of community structure on dynamic networks. Instead of indicating detailed changes of nodes and links temporally, BLOCKS regards communities as visual entities and focuses on representing their behaviors and relation changes on a time series. Experiments detected a stable performance of BLOCKS compared with previous methods while detecting the community structure of networks. We also present a case study that shows an effective learning process of network evolution with BLOCKS.