Directional-Layered Space-Time Densities: A New Spatiotemporal Trajectory Aggregation and Geographic Visualization Approach

One approach used in GIScience to investigate movement data is adopting time-geography theory and its important principles/methods: space-time paths and space-time cube. Thus, the movement of a moving object can be represented as a 3D polyline in a space-time cube. However, with the advent of larger movement datasets, this type of display can easily become cluttered and incomprehensible. In this article, we propose a new space-time aggregation algorithm, i.e., directional-layered space-time densities (DLSTDs), to solve the problem of trajectory clutter in a space-time cube. The approach is an extension of the pixel (2D raster cell) density around a point in two-dimensional space to the voxel density around a trajectory line segment in three-dimensional space-time $(x-y+t)$ . In the algorithm, the horizontal or vertical voxel layers are first calculated around the trajectory based on the direction of each trajectory line segment. Subsequently, the voxel density is calculated by substituting the distance of the voxel from the trajectory segment into a Gaussian-based attenuation function. Finally, the resulting layered densities are presented using volume rendering techniques. We demonstrate the DLSTDs on 30-day movement data of transport trucks in a mine in China and compare the algorithm with two other algorithms. At the end of the article, we summarize the research and consider further developments of the proposed approach.