University of Minnesota Driven to Discover
U of MNUniversity of Minnesota
Center for Transportation Studies

Visualization tools make sense of traffic data

crackedpavement

Four thousand vehicle detectors embedded in the pavement of the Twin Cities freeway system record traffic volumes, generating reams of data every day. In order to make sense of this ever-growing mountain of data, researchers need to create interactive graphical visualizations that allow them to see complex changes in traffic patterns. But with so much data to process, it can take days to create a visualization using traditional techniques.

To address the need for better, faster visualizations, Shashi Shekhar of the University’s Department of Computer Science and Engineering worked with engineers at Mn/DOT’s Regional Traffic Management Center (RTMC) to develop new high-performance techniques for visualizing traffic data. Shekhar is an expert in the area of spatial databases—databases designed to store spatially related information, such as that produced by the city-wide freeway sensor network.

With research assistants Chang-Tien Lu and Alan Liu, Shekhar analyzed the existing RTMC database and implemented advanced storage techniques to create a data warehouse. The warehouse aggregates certain types of data to facilitate visualization, and organizes information in a way that makes common queries more efficient.

The researchers also developed algorithms to discover which detectors are functioning abnormally. This is important from an operational perspective because it gives users the ability to identify sensors that may require repair or replacement, as well as exclude unreliable data when creating new traffic pattern visualizations.

One of the most strikingly useful visualization tools now available to RTMC engineers is the ability to create an animated map of traffic volume system-wide or on any freeway section at any time. Additional charting tools allow users to visualize and compare specific attributes of traffic flow.

The report authored by Shekhar and his research assistants: High-Performance Spatial Visualization of Traffic Data.