Retaining Interactivity in a Visual Analytics System for Massive Public Transportation Data Sets

Visual analytics aims to be scalable in several aspects, one of which is being able to handle large data volumes in an interactive system. This can be achieved by designing efficient analysis and visualization algorithms that ensure short response times to user interactions or by providing appropriate hardware. With increasing data sizes, however, an analyst will eventually have to wait for the completion of computations or renderings. For these cases, we present a framework which remains responsive and keeps the analyst informed on and in control of pending operations. Our implementation builds on performing complex analysis and rendering tasks in the background, providing progress indications, displaying preliminary results, and allowing changes to task elements while continuing to evaluate others. We demonstrate the advantages of our approach for huge data sets by analyzing a real-world public transportation vehicle data set.