| Abstract |
Background
The growing availability of high-accuracy geolocation Big Data has transformed transportation planning and modeling. Data from smartphone applications, connected vehicles, and fleet telematics systems are increasingly used to generate planning inputs such as Origin-Destination (O-D) matrices. However, practitioners often assume that one data source can serve all planning needs, which can lead to inefficient analyses and misleading results.
Key Issues
There is no single Big Data source that consistently provides the most accurate or reliable results across all transportation applications. Differences in data collection methods, sample composition, spatial and temporal coverage, and trip characteristics lead to variation in the outputs derived from each source. Selecting an inappropriate dataset can result in gaps, bias, or wasted resources.
Objectives
This presentation aims to compare the strengths, limitations, and appropriate use cases of the three primary high-accuracy geolocation Big Data sources used in transportation planning: Location-Based Services (LBS) data from smartphone applications, Connected Vehicle (CV) data from vehicle OEMs, and telematics data from Telematics Service Providers (TSPs). The objective is to help practitioners align data selection with specific planning and modeling needs.
Methodology
The presentation focuses on the generation and application of O-D matrices derived from LBS, CV, and TSP data. A two-step framework is proposed: first, clearly define the planning or modeling objective; second, select the data source—or combination of sources—that best supports that objective. Technical differences between the datasets are examined and real-world planning examples are used to illustrate how these differences affect modeling outcomes.
Conclusions
Case studies in the presentation demonstrate that different transportation applications benefit from different data sources. Transit planning may be best supported through a combination of LBS and TSP data, long-distance travel models can use LBS data to supplement gaps in CV datasets, and freight analysis is often most effectively informed by TSP data. Examples include a Texas Department of Transportation and Texas Transportation Institute statewide planning model, a Cascadia Rail Travel Market project by Washington State Department of Transportation (WSDOT) that includes zones into Canada, and a Passenger Rail Master Plan for Province of Alberta. The presentation concludes that thoughtful, objective-driven data selection is essential for producing accurate, reliable, and actionable transportation planning results.
|
