An evaluation of web-based geovisualizations for different levels of abstraction and realism – what do users predict

Web-based geovisualizations are produced and served in various levels of abstraction (or realism) such as two-dimensional (2D) cartographic maps, aerial and satellite maps, shaded relief maps, three-dimensional (3D) objects integrated with 2D base maps, and digital globes with fully textured realistic 3D representations. All of these are necessary; but which one is best fitting for which task? When do we need the highest level of realism, and when can or should we use the highest level of abstraction? To contribute to tackling these large questions, we study a subset of non-expert tasks selected from task taxonomies in literature in relation to a subset of existing geovisualizations by means of two online user studies. In an online survey, users (n=106) responded to a list of tasks, where we ask them to predict which of the visualizations they think they would use for this task (thus we measure perceived preference). In a second survey, we give the users (n=245) a set of tasks to solve using one of the provided visualizations, thus we observe which level of abstraction/realism they will actually use and measure actual preference (choice) as well performance. In this paper, we report the results from the first survey.