Visualizing Evolving Trees

Evolving trees arise in many real-life scenarios from computer file systems and dynamic call graphs, to fake news propagation disease spread. Most layout algorithms for static do not work well an evolving setting (e.g., they are designed be stable between time steps). Dynamic graph better suited this task, although often introduce unnecessary edge crossings. With mind we propose two methods visualizing that guarantee no crossings, while optimizing (1) desired length realization, (2) compactness, (3) stability. We evaluate the new methods, along with five prior approaches (three dynamic), on real-world datasets using quantitative metrics: stress, stability, running time. The fully functional available github. (This was supported part by NSF grants CCF-1740858, CCF-1712119, DMS-1839274.)