How close the models are to the reality? Comparison of transit origin-destination estimates with automatic fare collection data

There is a consensus on the importance and value of automatic fare collection (AFC) data in analysing different aspects of public transport. As such combining other data sources such as the General Transit Feed Specification (GTFS) can greatly improve the quality of the analyses and ultimately provide a better understanding of public transport performance. This paper presents a methodology for data processing and analysis to acquire a public transport Origin Destination (OD) matrix. The case study uses a very large dataset on passenger boarding and alighting of all three transit modes, namely bus, rail and ferry, in South-East Queensland (SEQ). The OD trip matrices are estimated for both the AM and PM peak periods for five weekdays. Also, the estimated public transport demands for the same periods from the SEQ strategic transport model (SEQSTM) are employed. This approach enables not only the comparison of OD matrices over time to determine changes in travel patterns but also investigates the similarity between the demands from the SEQSTM procedure and those from AFC data. A number of statistical measures, namely R, GEH, %RMSE and an eigenvalue-based measure, are utilized to determine the level of similarity of these OD matrices. The results highlight the similarity of the trip pattern between four workdays (Mon-Thu). However, trip patterns on a Friday are slightly different from the other weekdays, particularly in the PM peak period. Also, the demand from SEQSTM for both time periods is not analogous to any of the AFC patterns.