Using Visual Stability to Support Search Efficiency and User Experience in Dynamic Graph Drawings

In this paper, we present a set of algorithms to compute a foresighted graph layout. We aim to produce a dynamic graph drawing that allows the users to track actors in a dynamic social network, supporting a satisfying user experience along with an efficient visual search and recognition of elements. The drawings produced by the algorithms were evaluated through a study, which combined the use of questionnaires, an eye-tracking device and a model we developed to measure the visual stability of dynamic graph drawings. The results suggest dynamic graph drawings achieving visual stability by placing multiple actors in the same position are confusing and are not efficient in the visual search. On the contrary, introducing a restriction to prevent too many actors to occupy the same position produces the desired drawing.