PCNET: Exploratory visual analysis of large-scale network

The enormous growth of data in the last decades led to big data challenge in the network security field. Traditional visual analysis method for large-scale network exploration is inadequate. Efficient methods for visual clutter reduction, network structure exploration and network behavior detection are needed. In this paper, we propose two methods: Enhanced Histogram Brush (EHB) and Flow-based Fast Newman (FFN) algorithm aim to assist the visual analysis task in large-scale network exploration. The EHB is a novel improvement in Parallel Coordinates to guide exploratory interactions especially for big data. The FFN algorithm can efficiently discover the network hierarchy and extremely reduce the visual clutter in the network layout. A visual analysis tool PCNET is designed and implemented on the basis of these two novel methods. PCNET is capable of visually analyzing vast amounts of network data. To better describe and demonstrate the usefulness and performance of PCNET, we utilize the ChinaVis2015 Challenge dataset as a case study.