Bikeshare Pro: A Tactical Planning Tool

Bikeshare Pro is an interactive web-based platform that provides network design and analysis capabilities using data-driven and model-driven analytics. The tool integrates demand prediction and network optimization engines to design systems based on various objectives. It streamlines decision-making related to optimizing station locations, dock capacity, and charger availability while assessing performance against goals such as ridership, equity, accessibility, and transit integration. The geospatial unit of analysis in Bikeshare Pro is a hexagon grid system commonly used in other bikeshare master plans with an adjustable radius of 200-300 m.

Ridership Prediction
The ridership prediction engine uses state-of-the-art machine learning techniques, including Graph Neural Networks and XG-Boost, to predict manual bike and e-bike flows based on land-use, sociodemographic, and station-specific features (e.g., number of docks) of both origin and destination points. When prior ridership data is unavailable for model testing, data from similar cities regarding population, size, and sociodemographic characteristics is used as a proxy to calibrate the prediction model.
The prediction model is sensitive to network properties like dock capacity and charger availability, which influence e-bike flows. By adjusting these properties, the tool predicts system performance in line with the chosen objectives and allows users to refine the design until the desired outcome is achieved.

Network Optimization
The optimization engine recommends locations for new stations based on the selected objective or a weighted combination of multiple objectives. The model lets users specify the desired number of new stations and adjusts locations based on the weighted objectives. These recommended locations are further evaluated based on nearby points of interest within 300m and a customizable general scoring system.

Level of Service Analysis
The level of service is defined by bike availability at different stations. The tool provides a probability distribution of bike availability at each station for each month. Availability probabilities are influenced by the number of docks, sociodemographic and land-use data, and relocation strategies.

Relocation Analytics
Micromobility fleets often experience imbalances in station-based bike sharing, where some stations have a deficit and others a surplus of bikes. This imbalance can lead to rejected demand when no bikes are available or inconvenience when all docks are full. Bikeshare Pro offers analytics for balanced station planning and implementing relocation strategies to redistribute the fleet, thereby improving the level of service.

Cost/Revenue Analysis
The tool provides a cost/revenue cash flow analysis for upcoming years. Revenue projections are based on the ridership prediction model for regular bikes and e-bikes, which have distance-based pricing. Costs are divided into the fixed costs of station deployment and the annual operational costs of dock maintenance. Additional costs, such as fleet acquisition, can be added through customization.

E-bikes and Charging Stations
The tool predicts e-bike ridership based on charger availability within the network. State-of-charge (SOC) analytics is also available when the SOC-charge curve is provided. The tool allows any station to be electrified to become a charging station, influencing e-bike ridership as charger availability is statistically shown to impact usage.

Transit Integration
Dynamic heatmaps show how well the bikeshare system integrates with public transportation, specifically measuring how many bikeshare stations are within walking distance of major transit stations. The heatmap identifies three categories of transit stations: those with 0, 1, or 2+ bikeshare stations nearby.

Road Network Flows
Predicted station-to-station flows are mapped onto the road network using navigation APIs prioritizing cycling infrastructure. Overlaying these flows onto bike lanes reveals their usage in different parts of the city.

Continuous Micromobility Planning

Many cities have traditionally followed a “phased” bike-share planning strategy, whereby every 3-5 years, a Request for Proposals is issued to engage consultancies in developing comprehensive micro-mobility master plans. A master plan is a strategic document that outlines a long-term vision and framework for a city's micro-mobility service, offering an expansion plan considering features such as equity, transit integration, and accessibility.

The proposed “continuous” methodology contrasts with the “phased” in that it offers an intuitive (easy to learn) data-driven web-based tactical planning tool for network planning. The tool's automated nature allows for generating quick and responsive results based on the analysis request.

Bikeshare Pro a) engages the planners and city staff more actively in the planning process, offering flexibility in terms of their desired level of involvement, b) allows for fast turnaround analysis as the tool is highly automated, c) offers a dynamic planning environment where the planner can interactively engage with the tool, and d) is data-driven and calibrated on historical and real-time datasets.

The tool is designed to complement and sometimes replace the network planning side of traditional master plans. Elements like public engagement, however, still require experienced consultancies trained in this exercise. In summary, the tool offers a wide range of "what-if" scenarios in the design process for responding to evolving transportation challenges and opportunities in a more agile manner than traditional master plans.

Maps and Filters

Map and filter bike-share stations and ridership data based on criteria such as station type, trip characteristics, ridership levels, etc. Like a dashboard, this process facilitates the analysis of system trends including ridership patterns, mode choices (between regular and electric bikes), and system growth over time. The integration with transit and cycling networks is also investigated here.

Add/Remove Stations and Adjust their Properties

Add or remove stations and adjust their properties, such as dock capacity and charger availability. Deployment year can also be adjusted for future growth planning.

Ridership Prediction

Uses machine learning models to predict both regular and electric bike flow between stations, adjusting predictions dynamically as station properties change or when new stations are added. Ridership prediction is useful in assessing the impacts of fleet size and composition changes and investigating the implementation of e-bike expansion programs.

Transit Integration Analysis

Visualize the integration efficiency between transit stops and bike-share stations, highlighting the proximity of bike-share stations to transit stations within walkable distances. The feature updates dynamically as new stations are added to continuously assess and explore the efficiency of transit integration.

Electrification Analysis

Find the ideal locations for charger deployment to maximize e-bike ridership or other sociodemographic impacts, such as accessibility to e-bikes. A dedicated prediction model for electric flows allows for separate analysis, enriching ridership analytics.

Station Location Optimization

Optimize new bikeshare station locations by prioritizing accessibility, equity, and transit integration. Incorporate ridership in the optimization so that the stations would be located to maximize total ridership .

Micro-level Analysis

Assesses the micro-level properties of bike-share stations and their proximity to points of interest and transit stops, aiming to improve the system's accessibility and attractiveness.

Analytics

Provides a dedicated panel for analysis, charts, and tables at the station, ridership, and network levels, updated dynamically. The bottom panel offers an overview of network stations, including their properties, feature distributions, ridership patterns, and new station data. 

Cost and Revenue Analysis

Estimates the cost and revenue trends of the designed system based on the provided inputs for maintenance, infrastructure costs, and pricing structure.

Reports and Data Management

Download data in different formats and generate customized PDF reports for comprehensive insights and system summaries. Additionally, the design can be saved and reloaded, facilitating seamless management of data and analysis settings.

Bikeshare Pro has been customized for three case studies of Toronto, Vancouver, and Ottawa. The ridership prediction models of Vancouver and Toronto have been calibrated from the ridership data provided by their open data portals.

Bikeshare Pro is designed to be scalable to other cities. Contact us below if interested in seeing a version of Bikeshare Pro for your city.