Assessing passenger feedback reliability in crowd-sourced measurement of transit ride quality

Transit ride quality is important for everyday commuters to keep using public transportation as a preferred mode of travel. Despite well paved and maintained roads, sophisticated buses and modern bus control and monitoring systems, bus passengers, especially standing passengers, are frequently subjected to braking and acceleration actions along the way and sway discomforts along the curves. Modern buses come with rather powerful braking systems that are sensitive to slightest pressure on brakes by drivers. Passenger injuries as well as complaints increase as such. When crowd sourcing via using smartphone apps is employed, anyone ranging from a child to senior citizen can access and register a complaint about a discomfort moment along bus rides. Filtering genuine complaints is a challenge. We study a basic case of a person registering discomfort in a bus ride using the smartphone app developed for this purpose. At the back-end, the app keeps collecting sensor data. Findings provide a basis for developing machine-learning algorithms for big-data analytics of passenger feedback at a large-scale. A proof-of-concept program written based on study findings was used to show how discomfort data can be authenticated and classified.