Generalizing OD Maps to Explore Multi-dimensional Geospatial Datasets

Understanding the mobility of entities in geospatial data is important to many fields, ranging from the social sciences to epidemiology, economics or air traffic control. Visualizing such entities can be challenging as it requires preserving both their explicit properties (spatial trajectories) and their implicit properties (abstract attributes of those trajectories). An existing technique called origin-destination maps preserves both explicit and implicit properties of datasets, using the spatial nesting technique. In this paper, we aim at generalizing this technique beyond an origins-and-destinations dataset (2-attribute datasets), to explore multi-dimensional datasets (N-attribute datasets) with the nesting approach. We present an abstraction framework - we call Gridify - and an interactive open-source tool implementing this framework using several levels of nested maps. We report on several case studies representative of the types of dimensions found in geospatial datasets (quantitative, temporal, discrete, boolean), showing the applicability of this approach to achieve visual exploratory analysis tasks in various application domains.