SI2-SSE: Collaborative Research: TrajAnalytics: A Cloud-Based Visual Analytics Software System to Advance Transportation Studies Using Emerging Urban Trajectory Data

点击翻译按钮获取中文摘要

Advanced technologies in sensing and computing have created urban trajectory datasets of humans and vehicles travelling over urban networks. Understanding and analyzing the large-scale, complex data reflecting city dynamics is of great importance to enhance both human lives and urban environments. Domain practitioners, researchers, and decision-makers need to manage, query and visualize such big and dynamics data to conduct both exploratory and analytical tasks. To support these tasks, this project develops a open source software tool, named TrajAnalytics, which integrates computational techniques of scalable trajectory database, intuitive and interactive visualization, and high-end computers, to explicitly facilitate interactive data analytics of the emerging trajectory data.The software provides a new platform for researchers in transportation assessment and planning to directly study moving populations in urban spaces. Researchers in social, economic and behavior sciences can use the software to understand the complicated mechanisms of urban security, economic activities, behavior trends, and so on. Specifically, this software advances the research activities in the community of transportation studies. The software also acts as an outreach platform allowing government agencies to communicate more effectively with the public with the real-world dynamics of city traffic, vehicles, and networks. The integration of research and education prepares the next generation workforce, where students across multi-disciplines can benefit through their participation in the development and use of the software.In order to lay the foundations for effective analysis of urban trajectory datasets, this software (1) facilitates easy access and unprecedented capability for researchers and analysts with a cloud-based storage and computing infrastructure; (2) develops a parallel graph based data model, named TrajGraph, where massive trajectories are managed on a large-scale graph created from urban networks which is naturally amenable for studying urban dynamics; and (3) provides a visualization interface, named TrajVis, which supports interactive visual analytics tasks with a set of visualization tools and interaction functions. A variety of transportation-related researchers from geography, civil engineering, computer science participate in the design, evaluation, and use of the software. This robust and easy-to-use software enables the users to conduct iterative, evolving information foraging and sense making.