Lead Presenter Biography

Kyeongsu Kim is a passionate transportation data analysis expert. He has over 16 years of experience in data analytics and modeling. Kyeongsu has led numerous innovative data analytics and modeling projects and has a proven track record of exceeding client expectations. He is especially interested in data-driven techniques that leverage surveys, census, housing, land use and other passively and actively-collected data to better understand meso- and macro-level travel behavior. His analytical and project management skills help bridge the project delivery gap between consultants and clients.

Prior to RSG, Kyeongsu served as a project manager at Connetics Transportation Group (now Nelson/Nygaard) working with local and regional clients/consultants and leading and supporting travel demand modeling, application, and data analytics projects/tasks. Prior to CTG, he supported research and data analysis efforts at New York Metropolitan Transportation Council while working for Louis Berger (now WSP).

Kyeongsu majored in transportation and logistics management at Korea Aerospace University, received a master’s degree in City and Regional Planning from Rutgers University and a PhD in geography from the Graduate Center of the City University of New York. He lives with his wife and 8-years old son in Fairfax, Virginia. He enjoys playing baseball with his son who wants to go pro one day.

Co-Authors

Presentation Description

Interest in visualizers/dashboards to illustrate the inputs and outputs of travel demand models has grown over the past few years among modelers, planners, and policymakers. There are slightly different needs by each group. Typically, modelers look for tools that facilitate model validation and post-processing with minimal manual data wrangling. Planners and policymakers, on the other hands, seek dashboard-type tools that expedite transportation planning and policymaking process by transforming complex model inputs/outputs into simplified, interactively readable maps, tables, charts, and narratives.

RSG has developed many visualizers and dashboards to meet those needs either for internal purposes or as a project deliverable. We developed another internal visualizer, using a python library, Dash, for our recent project with Fredericksburg Area Metropolitan Planning Organization (FAMPO).

The project was to examine challenges in the overall east-west traffic movement of the Fredericksburg Metropolitan Area where most economic activities have been concentrated along the north-south I-95. The project team proposed 15 roadway improvement scenarios to assess their impacts on congestion relief, and each of the projects was being coded into the highway network and other related files in the CUBE environment where the FAMPO model was built. After each scenario run, CUBE generated a highway network loaded with the speed and volume of each road link, and that information are translated into project performance measures such as Volume/Capacity Ratio (V/C Ratio), Vehicle-Hours-Traveled (VHT), Vehicle-Miles-Traveled (VMT), etc. The project team adopted a congestion severity concept based on V/C ratio and Percent of Congested Travel (PCT). Road links were categorized into four congestion types (extreme, severe, moderate, and free) based on facility type and V/C ratio. PCT was calculated by dividing congested VMT by total VMT. At the regional level, the PCT was used as a key measure of whether a project benefits the entire region.

These performance measures are illustrated in the visualizer. For example, the visualizer has a color-coded map showing links with congestion severity and offers a drop-down menu to select specific scenario run results. We used the visualizer to evaluate the results of each scenario run, compare performance measures across multiple scenarios, and facilitate the entire post-processing of model test runs.

Visualizer is directly connected to CUBE’s scenario directory in a local machine. When the model run of a proposed project is complete, the visualizer automatically adds the scenario to a list of available scenarios to show. Rendered maps and statistics are ready for users to view and compare with different scenarios. Visualizer’s front end is designed to be user-friendly and feature-rich for planners and policymakers to explore. Users can select a scenario of interest from a dropdown list and upon selection, the map and the defined performance measures will be updated. By clicking on the road segment of interest on the map, users can see a table of road segments with details on V/C Ratios, VMT, and VHT. Maps and performance measures are shown alongside the base scenario to illustrate the improvement. This visualizer can be applied to other regions without any heavy lifting. 

Presentation File