Pedal preferences: GPS-based panel data insights into bike share traffic flow across membership groups

Bike share systems promote sustainable transportation and active mobility. Understanding spatiotemporal usage patterns and influencing factors is crucial for equitable and effective policies. This study analyzes a full year of Global Positioning System-tracked bike share trip data from Hamilton, Ontario, to examine the travel behaviors of three membership types: Monthly and Seasonal Members, Pay-As-You-Go riders, and McMaster Monthly Pass holders. We employ descriptive statistics to analyze trip start times and the most frequently used routes, alongside a two-way fixed-effect binary logistic model to investigate bike share traffic at the road-and-day level, providing detailed insights into the determinants of bike share usage. Findings reveal that different membership types exhibit distinct spatiotemporal usage patterns and preferences regarding land use, infrastructure, sociodemographics, and events affecting bike-share road traffic. Only Monthly and Seasonal Members display consistent commuting patterns throughout the year. McMaster Monthly Pass holders dominate during the school semester following the introduction of a discounted pass for undergraduate students. Furthermore, Monthly and Seasonal Members are more likely to cycle on roads adjacent to parks, while McMaster Monthly Pass holders show lower sensitivity to extreme temperatures. Precipitation, darkness, slope, and holidays consistently deter bike share usage. Policy recommendations include expanding fare discount programs, improving wayfinding, organizing cycling events during holidays, and enhancing winter road maintenance for heavily used cycling routes. This study highlights differences in usage patterns, distinct preferences, and varying sensitivities to factors affecting bike share traffic flow among membership types, offering robust insights through a long study period and detailed road-level data. • Analyzed 440,173 GPS bike share trips over a full calendar year. • Regression was used to model bike share traffic presence at the road-and-day level. • Bike share usage patterns differ across membership groups. • Land use, infrastructure, sociodemographics, and events affect bike share activity. • Precipitation, darkness, slope, and holidays consistently deter bike share usage.