Transit agencies depend on rider surveys to understand who their customers are and where they are traveling. Data derived from on-board rider surveys is used to guide planning decisions and conduct equity analyses critical to implementing transit system enhancements in a strategic and balanced way. Agencies could also be applying this data in more advanced ways, but only if the data is high quality and accurate.
Data In, Data Out
Foursquare ITP works with our agency partners throughout the country to envision the future of transit, using survey data to model rider behavior and prioritize transit improvements. However, after evaluating thousands of survey responses over the years a consistent pattern began to emerge from rider survey responses — oftentimes riders, either by mistake or purposefully, failed to enter correct information on routes and destinations. To address this, we asked our Data Science Team to develop a method to improve the overall quality of gathered data and to correct the problem of mistakes in survey location data. In response, the team created the Trip Chain Validator tool which validates whether travel patterns and routes reported in an on-board survey are indeed feasible on a given agency’s transit system. For example, if a person reports that they traveled on a route to a particular shopping mall—but that route doesn’t go there—the survey response would be flagged for review.
The Trip Chain Validator provides both feedback on survey data quality and enhances survey outputs. For tablet-based on-board surveys, validation can occur in near real-time to ensure accurate responses are included in a sample. The validation process also infers other aspects of a person’s trip, such as walk time to transit or likely transfer locations.
Foursquare ITP developed the tool with the help of insights from our partner firm WBA Research, which conducts on-board rider surveys. This valuable tool has already been put to use on projects for the Niagara Frontier Transportation Authority (NFTA) in Buffalo, NY and the Jacksonville Transportation Authority (JTA) in Jacksonville, FL. The Trip Chain Validator tool helped to improve the quality of location data in the Buffalo survey by more than 60%. “This tool ensures the data we collect is accurate and gives our clients confidence that they’re receiving a true picture of their transit system. With near real-time feedback on responses, it allows us to edit and process data faster. This tool allows agencies to get a more vivid picture of how people use transit,” said Kevin Pullis, Vice President with WBA Research. The tool has also been successfully used to help the Maryland Transit Administration identify top locations for bus transfers within its system to aid in prioritizing infrastructure improvements.