Identifying public transit deserts: A travel demand-independent persistent homology-based method

To identify time-space locations where public transit infrastructure fails to provide a reliable and timely alternative to private vehicles, this paper proposes a new travel demand-independent persistent homology-based method to locate and rank the severity of modal travel desert. Persistent homology, which is a tool from algebraic topology, is incorporated and the severity of a transit desert is measured as a trade-off between its proximity to existing transit infrastructure and the travel time required to travel through it. The proposed method highlights entire regions of cities that are bereft of suitable public transit, providing reasonable estimates even in the absence of travel demand data. This paper presents the techniques and software tools used to study the Stockholm public transit network. The proposed method is potentially useful for city planners to consider the trade-off between how severe a bottleneck is and how difficult the bottleneck is to fix.