Modelling Spatio-Temporal Variability in Informally Run Transport Routes to Improve Journey Planning Calculations

Abstract

Initiatives to map and digitalise informally run transport systems such as minibus taxis are becoming increasingly common in cities across the world (e.g. Digital Matatus, Transport for Cairo, Bus Map Beirut). Many of these projects collect the basic information on the informal routes needed to form a General Transit Feed Specification (GTFS) file, a digital method created to represent scheduled, fixed route networks largely used for the purposes of journey planning. However, a single collected trip on a route would fail to capture the high variability in even a single informally run transport route. Routes and vehicle frequencies fluctuate due to factors such as time of day, vehicle fullness, specialised requests, and route disruptions. Single route collections result in GTFS files that inaccurately translate real world dynamics into a single static piece of information. Furthermore, these GTFS files are fed into journey planning platforms which generate calculations to determine different multimodal route options from a given origin to destination. Without a better understanding of the factors that affect route geometry, the ability to better model real world informally run transport operations is limited, thereby affecting the accuracy of journey planning. The objective of this study is to determine the spatio-temporal variability of a single minibus taxi route in Cape Town, South Africa to inform how real world dynamics should translate into digital journey planning calculations to deliver more relevant information to users. This study finds through on-board data collection and roadside frequency counts that deviations in route geometry, fluctuations in trip duration, variance in route frequency, and average fullness factor along a route all bear implications journey planning tools