Dynamic visualization variables in animation to support monitoring of spatial phenomena
My research attempted to shed light on some animation design and use issues. From the design point of view, animations – being dynamic by nature – seem particularly suitable to quickly obtain qualitative information from large temporal data sets, like series of remotely sensed images. But animations have limitations as well, they are overwhelming and problems like ‘change blindness’ may occur. Little is known about how animations are actually used. Is it possible to extract relevant information from an animation? The interactive prototype aNimVis was produced to evaluate animation design and use issues in a monitoring context. The prototype was limited to interaction tools that affect the variables of the temporal dimension of an animation: ‘moment of display’, ‘order’, ‘duration’ and ‘frequency’. As a case study, satellite data containing a commonly used vegetation index (NDVI) were used and experts in monitoring participated in the evaluation.
The study revealed, amongst others, three different animation use strategies. Users made frequent temporal selections. Surprisingly given the potentially overwhelming character of an animation, they all used ‘tuning’, the mode to compare two simultaneously displayed animations and they made similar recommendations to further extend the functionality of this mode. The users wanted to play the animation to view the ‘dynamic behaviour’ and the ‘rate of change’ (two effects of the dynamic visualization variables). Other frequently generated effects by user’s interaction were ‘review’, ‘visual isolation’ and, in case of comparisons in time, ‘synchronization’. Not all tools provided in aNimVis were successful but overall, high usability rates were given and all users liked to integrate this visual exploration environment with a computational environment (e.g. a GIS or image processing software). If we know what effects users want, new tools can be designed that generate these effects.
Further tool development for animations should concentrate on tools to visually isolate part of the complex content of an animated representation.