Methods of determining the O-D matrix in the absence of data surveys

Abstract

Public transport is like a living organism that is constantly changing along with the city and the needs of its residents. Based on the changes in the city, its infrastructure or the habits of its residents, the public transport system must also change. One of the main tools for understanding the public transport in urban areas is passenger origin-destination (OD) matrices. OD matrices are compiled based on mass population surveys, during which the city is divided into zones, and the collected data is used to create travel patterns. Such studies implementation is expensive and repeated infrequently – usually every few or even ten years. Currently, on-board computers collect a significant amount of data about daily citizens’ trips, but the main problem is that this data is not enough to create an O-D matrix, because only boarding data is recorded. The second largest city in Lithuania was chosen for the study due to the abundance of data collected. The aim of this study is to develop a methodology based only on boarding data, allowing to reflect the passenger origin-destination matrix. Two methods were later tested during the study: the first, matching sequential entries based on that the same anonymised card identifier, and the second, the time that the passenger gets off at the stop where they get back on. The results of these methods are analyzed and evaluated at the stop and zone level.