Project Description

A common task involved in the monitoring of public transport services is the measurement of passenger boardings and the estimation of distances travelled by those passengers. This is often done by means of the conduct of sample surveys of passenger boardings and alightings and the calculation of the average distance travelled by passengers of various types.

This project covers the development and testing of a different method of calculating average distance travelled on a bus route that uses data obtained from automatic ticketing machines rather than from sample surveys. Importantly, the method relies only on the locations at which passengers board, and does not need to know where each passenger alights. Rather, a simplifying assumption is made that most bus trips are round trips and that passengers board the bus for their return journey at about the same place as where they got off on their forward journey.

The method has been tested against trip lengths estimated from a comprehensive origin-destination survey carried out on National Bus Company buses in Melbourne, Australia in April-May 1994, and has been shown to give very good estimates of average distance travelled. The error in estimating total distance travelled by adult full-fare and adult concession passengers is less than 0.5%, while the error in estimating total distance travelled by student passengers is about 3%. The larger error for students is primarily due to the smaller passenger flows for students.

Given these results, it is concluded that the "up-down" method of calculating average distance travelled based on boarding data obtained from automatic counts provides a better, cheaper and more flexible method of calculating average distance travelled than using sample surveys which measure boarding and alighting location.