Multi-similarity matrices of eye movement data

We describe a matrix-based visualization technique for algorithmically and visually comparing metrics in eye movement data. To reach this goal, a set of scanpath trajectories is first preprocessed and transformed into a set of metrics describing commonalities and differences of eye movement trajectories. To keep the generated diagrams simple, understandable, and free of visual clutter we visually encode the generated dataset into the cells of a matrix. Apart from just incorporating one individual metric of the dataset into a matrix cell, we extend this standard visualization by a dimensionalstacking approach supporting the display of several of those metrics integrated into one matrix cell. To further improve the readability and pattern finding among those values, our approach supports a metric-based clustering and further interaction techniques to manipulate the data and to navigate in it. To illustrate the usefulness of the system, we applied it to an eye movement dataset about the reading behavior of metro maps. Finally, we discuss limitations and scalability issues of the approach.