Interactive Visualization of Massive Location Data using Multi-Scale Trajectories

Abstract

Today location information on aircraft and vessels is collected worldwide and readily available to analysts. When analyzing such data for a large area or a long period of time, the sheer size of the dataset becomes a challenge. Especially when one wants to work interactively at both overview and detail scales. We present a scalable approach to visualize such data by treating it as a set of trajectories simplified at different error rates. When combined with tiling and GPU-based visualization, we can interactively and visually analyze a dataset of 1 billion position records on a workstation.