Abstract Manager

Integrating bikeshare systems into broader transportation networks is a key strategy in an increasing demand for active transportation, where sustainable and efficient mobility is paramount. Accurately predicting bike-share trip volumes in station-based systems can improve station location efficiency. By adding new bike-share stations strategically, the transportation network can cover a larger area and offer improved first- and last-mile access to public transportation. Additionally, planners need to interactively investigate various expansion scenarios, including adding new stations or dock capacity. Bike-share growth plans, typically carried out every 4-5 years, mean that managers have limited flexibility to deviate from these recommendations unless further studies. To bridge these gaps, we have introduced an advanced GIS-based tool for strategic bike-sharing planning, enabling manipulation of the recommendations while evaluating the impacts of alternative designs.

This study relies on data analytics and incorporates a deep machine-learning model to predict bike-share station transactions. This approach streamlines the process and reduces the need for flow assignment, leading to more efficient and effective management. Moreover, our study's facility location optimization function represents a multi-objective optimization, allowing the solution to be adjusted based on the weights assigned to each objective. We consider four primary objectives in selecting new station locations: accessibility, equity, transit integration, and ridership. The tool developed in this study is interactive, data-driven, intuitive, and accessible on the cloud. It offers bike flow predictions for potential new stations, analyzes equity based on city’s definitions and proximity to points of interest, investigates transit integration efficiency, optimizes station location, and supports the expansion of e-bikes.

The primary case study for the tool was conducted on the Toronto bike-share network. We sourced the primary data from the City of Toronto's open data portal. Many cities offer similar data, allowing the tool to be adapted for various providers. The prediction model within the tool considers factors like demographic information and neighbouring stations and their proximity to public transportation. The facility location optimization model uses Toronto's Neighbourhood Equity Index, population, and transit integration based on the closeness of bike-share stations to transit stations. These approaches are seamlessly integrated into the platform, enabling analysts to evaluate bike-share flow between each pair of stations when modifying the quantity and features of stations, or adding and removing them. Our goal is to transform bike-sharing network management with a data-driven solution that is efficient in its operation and equitable in design while considering the entire multimodal transportation network.