Station and city level modelling of bike-sharing system for Montreal

As an emerging mode of sustainable urban mobility, bike-sharing systems (BSS) have prevailed in North America in the past decade. As a result, there is a growing need to design and plan BSS to fit the unique urban transportation system of each city. This task, however, requires a deep understanding of how diverse factors such as built environment, land use, and weather conditions affect cycling behavior. To understand BSS user's general behavior with Montreal's weather condition, we first looked at the total number of BSS cyclists of the city. However, while some of the factors can impact the entire city, others can affect the riders locally. Considering the local impact of the factors can further help planners to realize the possible deficiencies more fundamentally. Thus, this study was performed on the bike-sharing ridership in two different levels of analysis:In the first section, we performed a city-level investigation of how weather variables affect the cycling behavior of BSS users. In particular, we perform three regression analyses to understand how various weather variables such as temperature, rain, and humidity interact with the daily ridership in Montreal using the trip data provided by BIXI. The average daily temperature was the most influential weather factor on the number of cyclists. Additionally, a higher level of temperature elasticity was found for the trips on the weekends than on the weekdays. The city-level analysis was then projected to the next four decades to study the possible effect of climate change on cycling behavior as a common physical activity. This was achieved by using the developed regression models and future climate data obtained from state-of-the-art regional climate models for two emission scenarios. Results suggest a general increase in the number of users, which is particularly prominent for the shoulder months of April and October and is primarily due to the future warmer temperatures.The second part of this thesis focuses on further understanding how the built environment, land use, and weather factors jointly affect cycling behavior. In doing so, a spatiotemporally weighted regression model was developed for the daily ridership data at station level. This is achieved by integrating weather variables into the temporal dimension and built environment and land-use related variables in the spatial dimension. The spatiotemporal regression model enabled tracking the influence of each factor through both time and space. Spatial factors like parks, bike lanes, and commercial places demonstrated a more positive effect on weekends throughout the year. On the other hand, factors such as closeness to metro stations, walkscore, and the capacity of bicycle stations, had a more positive impact on weekdays. As for the weather variables, temperature was also found to have a high positive effect throughout the year, with a higher impact on the weekends and recreational stations of Old Port, Montreal Olympic Park, and Jeanne-Mance park. Furthermore, in this study, we also expanded on the previous modelling with spatial-varying coefficients. By including the temporal variables of weather parameters over the period of consideration, the spatiotemporal model contributes to a more robust regression model for predicting station ridership

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.