- November 7, 2018
- Perspectives
Using Big Data to Improve Transportation Modeling
Anthony Gallo, P.E.
Transportation Engineer
Tim Padgett, P.E.
Transportation Engineer
Anthony Gallo, P.E.
Transportation Engineer
Tim Padgett, P.E.
Transportation Engineer
Big data—such as probe data from mobile devices—has been a major development for the transportation planning and operations field. It’s now possible to conduct straightforward analyses that would have been challenging, expensive, or time-consuming even a few years ago. At Kimley-Horn, we maintain relationships with a variety of big data vendors, staying up-to-date on the latest in data analytics. Kimley-Horn’s Anthony Gallo, P.E. and Tim Padgett, P.E. recently presented on their use of one big data vendor—StreetLight Data—at TRB’s Tools of the Trade Transportation Planning Conference in Kansas City, MO. In this article, they share how big data can be used to improve transportation modeling.
What Is Big Data?
“Big data” is a catch-all term referring to large, complex data sets from new data sources that are so voluminous that traditional data processing software—for example, Microsoft Excel—cannot process all the information. The big data described throughout this article— and used in many transportation planning applications—comes from anonymized mobile device data created by cellphones (from location-based apps or GPS) and connected vehicles that use on-board navigation tools like TomTom. The data is collected by sampling many of these devices, and then using tools to process, aggregate, and anonymize the information gathered. StreetLight Data, the vendor in this example, provides aggregated metrics based on GPS (navigation) data and location-based services data. Big data is a new paradigm for transportation analyses, offering insight into speed/travel times, origin-destination paths, trip purposes, vehicle types, and more.
What Are Some of the Applications of Big Data?
Understanding Cut-Through and Diverting Traffic
Big data is a valuable tool for studying cut-through traffic through downtown areas or specific intersections and segments. Kimley-Horn recently used StreetLight Data to examine the amount of cut-through traffic along origin-destination (O-D) pairs in Hoboken, NJ. The city has limited entry/exit points, is adjacent to tunnels to New York City, and contains the Hoboken Terminal—a hub for travel by train, bus, and ferry. StreetLight Data allowed Kimley-Horn to quantify the amount of cut-through traffic versus local traffic, finding that more than one-third of the personal car trips and more than half of the truck trips that enter Hoboken on an average weekday are cut-through. Data analysis also helped identify specific O-D pairs that favored cut-through traffic. This information can be used for planning applications, such as determining who should be involved in the planning/development of a bypass road, and also for travel demand management, such as identifying weight restrictions or access restrictions for certain roads and areas.
Kimley-Horn also used the Hoboken data sets to conduct several supplemental data analyses, including travel times along a highway before and after a road diet, an O-D analysis for Hoboken Terminal, an O-D analysis for a waterfront roadway, and an O-D analysis for municipal parking garages. Kimley-Horn has also used StreetLight Data to quickly conduct cut-through analyses in several other locations nationwide, including Colorado, Texas, and Virginia.
Developing and Validating Models
Estimations of regional origins and destinations can be made using big data. These estimations can then be compared to those created with data from surveys and other methods to validate O-D data along specific corridors, or demand within a traffic model. Kimley-Horn recently used StreetLight Data in Greenville, SC to help inform a corridor study and validate an existing external station matrix for a regional travel demand model. For the corridor study, StreetLight Data was used to determine how much traffic was destined to the corridor versus how much traffic was just passing through. For the external station matrix, using StreetLight Data allowed for comparative analysis of the highest value stations, station magnitudes, and specific O-D interchanges.
What Should Be Kept in Mind When Using Big Data?
Big data has provided traffic planners with new tools for customized and sophisticated transportation analyses. With these new tools come caveats that are continuously being worked on and improved. For instance, are we capturing all modes of transportation well? Are sample sizes in smaller/less urbanized areas sufficient for thorough analyses? More research and validation needs to occur before we can consider big data as a comprehensive solution for planning challenges.
Kimley-Horn is well-connected with big data providers across the transportation industry. The analyses Kimley-Horn provides have a wide range of uses from traffic and transit planning to site development applications. Our localized knowledge coupled with our nationwide expertise allows Kimley-Horn teammates to understand and contextualize the insights and limitations of big data, converting large amounts of data into digestible, valuable information for clients.
About the Authors