NetFork: Mapping Time to Space in Network Visualization

Dynamic network visualization aims at representing the evolution of relational information in a readable, scalable, and effective way. A natural approach, called 'time-to-time mapping', consists of computing a representation of the network at each time step and animating the transition between subsequent time steps. However, recent literature recommends to represent time-related events by means of static graphic counterparts, realizing the so called 'time-to-space mapping'. This paradigm has been successfully applied to networks where nodes and edges are subject to a restricted set of events: appearances, disappearances, and attribute changes. In this paper we describe NetFork, a system that conveys the timings and the impact of path changes that occur in a routing network by suitable time-to-space metaphors, without relying on the time-to-time mapping adopted by the play-back interfaces of alternative network monitoring tools. A user study and a comparison with the state of the art show that users can leverage on high level static representations to quickly assess the quantity and quality of the path dynamics that took place in the network.