Incorporating Public Transit into Measures of Accessibility
Student Capstone Project
In the past, Statistics Canada developed network accessibility measures based on the distance of driving and walking to compute proximity scores for various types of amenities. However, the importance of public transit as a primary mode of travel has not been included and accessibility measures based on time using transit have never been incorporated into proximity scores. This is why Statistics Canada enlisted the help of a team of students from UBC’s Master of Data Science Okanagan program to help develop a high-performing, efficient and scalable model for computing and visualizing public transit accessibility across Canadian metropolitan areas.
Transportation network analysis is vital to urban planners, researchers, and policy makers to ensure accessibility is considered when determining how resources are distributed across a population.
This requires both a realistic and scalable measure for accessibility, which can impact the planning and development of roads and city districts, while also helping to analyze a population’s access to healthcare, schools, grocery stores, and other amenities.
In the past, Statistics Canada, the government agency responsible for producing data to better understand Canada’s population, resources, economy, society, and culture, developed network accessibility measures based on the distance of driving and walking to compute proximity scores for various types of amenities.
But the importance of public transit as a primary mode of travel has not been included and accessibility measures based on time using transit have never been incorporated into proximity scores due to its complexity and computational intensity.
As such, Statistics Canada enlisted the help of a team of students from the University of British Columbia’s Master of Data Science Okanagan program to help develop a high-performing, efficient and scalable model for computing and visualizing public transit accessibility across Canadian metropolitan areas.
The goal of the capstone project was to develop a model to determine how accessible cultural amenities are in Vancouver, based on the transit routes within Metro Vancouver for different days of week and times of day.
Over the course of the capstone project, UBC’s MDS-O students used Conveyal’s R5 realistic routing to address the computational intensity of multimodal shortest path routing. By computing over 5.3 million transit routes repeatedly, 360 times in a day over three days, equating to roughly 5.7 billion shortest path searches in under one hour, the project showed other metropolitan areas in Canada could be similarly assessed without facing large computational hurdles.
The capstone project also detailed the advantages, disadvantages, and possible use cases of three unique accessibility measures, all of which are calculated from the travel time matrix. The project illustrated how these measures can be used for analysis both visually, and statistically using various techniques such as k-means clustering and Pearson's correlation coefficient matrices. The developed accessibility measures were leveraged to probe for new ways of understanding network efficiency and the travel time matrix was then used to develop three accessibility measures: one based on time, one on scores, and one on percentiles which were visualized with Leaflet and Kepler.gl and embedded in an R shiny dashboard.
This project provided a high performing and scalable framework for producing three unique transit accessibility measures for network analysis using Greater Vancouver as an initial use-case scenario. The frameworks can be further developed and adopted by urban developers to ensure equitable, sustainable, and optimal urban design for years to come. Most importantly, this framework can be scaled across new amenity types, including schools, hospitals, parks and grocery stores, and across other metropolitan areas in Canada to provide a basic methodology for future Census studies regarding transit accessibility measures. This framework provides a basis for future analysis and development that can be adopted by city planners and policy makers to ensure equitable, sustainable, and optimal urban growth for population segments which depend on public transit.