Visualizing spreading phenomena on complex networks (1807.01390v1)

Abstract: Graph drawings are useful tools for exploring the structure and dynamics of data that can be represented by pair-wise relationships among a set of objects. Typical real-world social, biological or technological networks exhibit high complexity resulting from a large number and broad heterogeneity of objects and relationships. Thus, mapping these networks into a low-dimensional space to visualize the dynamics of network-driven processes is a challenging task. Often we want to analyze how a single node is influenced by or is influencing its local network as the source of a spreading process. Here I present a network layout algorithm for graphs with millions of nodes that visualizes spreading phenomena from the perspective of a single node. The algorithm consists of three stages to allow for an interactive graph exploration: First, a global solution for the network layout is found in spherical space that minimizes distance errors between all nodes. Second, a focal node is interactively selected, and distances to this node are further optimized. Third, node coordinates are mapped to a circular representation and drawn with additional features to represent the network-driven phenomenon. The effectiveness and scalability of this method are shown for a large collaboration network of scientists, where we are interested in the citation dynamics around a focal author.