Graph Analysis and Visualization: Discovering Business Opportunity in Linked Data

Chapter 15 Dynamic Graphs

Dynamic graphs represent data where nodes and links are created or removed from one point to another. Some such graphs are only concerned with viewing a snapshot of the graph at a single point, or collapse time to include all nodes and links from start to end. This time focuses on the other cases, when change in time or scenario is an important dimension of the story being portrayed.

Representing a time dimension can be particularly challenging with graphs. In scientific circles it is generally accepted that simultaneous representations of states over time are more informative than sequential representations. For example, a time series bar chart is much better at communicating behavior over time than animating changes to a single bar. In the latter case, scrolling the animation back and forth through time repeatedly would be a means of getting the gist of change but would not be as instantly accessible as a time series, as accurate for comparison, or as easy to spot correlations in behavior. Being able to see values over time at the same time is more effective. However, it may not be immediately obvious how to apply that principle to graphs.

In a graph, like a map, the horizontal and vertical dimensions are reserved for the fundamental representation itself, so they cannot be used to portray time. This presents a challenge when using both types of representations. However, things get even more challenging when using graphs. The spatial location of ...