Contents lists available at ScienceDirect Travel Behaviour and Society journal homepage: www.elsevier.com/locate/tbs What factors influence bike share ridership? An investigation of Hamilton, Ontario’s bike share hubs Darren M. Scott ⁎ , Celenna Ciuro TransLAB (Transportation Research Lab), School of Geography and Earth Sciences, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4K1, Canada ARTICLE INFO Keywords: Active travel Bicycle Bike share Cycling Hub Ridership Travel behavior ABSTRACT Hamilton, Ontario’s bike share system was launched officially on March 22, 2015. This study analyzes the effects of weather conditions, temporal variables, hub attributes (most of which are derived for 200 m buffers around hubs), and a one-day lag on daily ridership at the bike share’s hubs during its first year of operation. Two random intercept multilevel models are estimated – one for daily trip departures, the other for daily trip arrivals. All weather (temperature, precipitation) and temporal variables (daylight hours, university terms, weekdays, holidays) are statistically significant in both models. Conversely, variables measuring transportation infra- structure in the vicinity of hubs, including the amount of bike lanes, are largely insignificant, suggesting that these features of the built environment have little to no influence on ridership. Proximity to important locations in the city (McMaster University, Hamilton’s downtown) has a strong impact on ridership. Although population density was an important consideration when locating the hubs, population does not influence daily departures or arrivals. Employment in the vicinity of hubs, which serves as a surrogate for an area’s activities or its at- tractiveness, does influence ridership, as is the case for the one-day lag effect. While all of these variables are able to explain some of the differences in daily ridership activity between hubs, the random intercept models confirm that they do not explain all of it. In other words, there remain intrinsic differences between hubs that are not captured by the independent variables – differences that influence ridership. 1. Introduction In recent years, bike share systems have gained in popularity as they have come to serve as an increasingly convenient source of transport (Fishman, 2016; Shaheen et al., 2013). From just a handful of bike share systems operating in the late 1990s, this list has grown to 1950 worldwide as of March 2019 with almost 383 systems more being planned or under construction (Meddin, 2019). Bike shares provide an alternative, sustainable mode of transportation contributing to the creation of healthy cities. There are multiple benefits of bike share systems including, but not limited to, flexible mobility, reduced emis- sions, health benefits, reduced congestion and fuel use, individual fi- nancial savings, and support for multimodal transport connections (Shaheen et al., 2013). With respect to the latter benefit, bike shares enhance access to and from public transit, thus improving upon the issue of last-mile connectivity (Jäppinen et al., 2013). Bike shares also normalize the image of cycling as an everyday travel mode, thus broadening the cycling demographic (Goodman et al., 2014). Bike shares not only encourage individuals to become more environmentally conscious, they promote active transportation that can enhance physical activity levels to obtain better health outcomes (Faghih-Imani and Eluru, 2015). While the number of studies investigating bike share ridership in North American cities has increased in recent years, findings con- cerning factors hypothesized to influence ridership are extremely dif- ficult to compare across studies due to methodological differences and contextual differences. A city’s geographical context, its site, de- termines many factors that are hypothesized to influence bike share ridership, including its weather, seasons, and daylight hours. Further, the location of a bike share within a city determines its usage – not only absolute ridership, but also who uses it and when. For example, it is likely that ridership in many of the large North American cities that have been studied by researchers (e.g., Chicago – Faghih-Imani and Eluru, 2015; Montreal – Faghih-Imani et al., 2014; Toronto – El-Assi et al., 2017; Washington D.C. – Gebhart and Noland, 2014) is affected not only by local users, but also tourists. On the other hand, ridership in small to mid-size cities that are not typically major tourist destinations is mostly impacted by local residents using their systems. Contextual differences alone suggest that our understanding of factors influencing bike share usage can benefit from further studies in different cities such https://doi.org/10.1016/j.tbs.2019.04.003 Received 14 February 2018; Received in revised form 14 July 2018; Accepted 8 April 2019 ⁎ Corresponding author. E-mail addresses: scottdm@mcmaster.ca (D.M. Scott), ciurocj@mcmaster.ca (C. Ciuro). Travel Behaviour and Society 16 (2019) 50–58 2214-367X/ © 2019 Hong Kong Society for Transportation Studies. Published by Elsevier Ltd. All rights reserved. T