Empirical evidence on Transfer Time and Distance in a Public Transport Network Using Smart Card Fare Data

Over the past few years, a number of techniques have been developed to estimate public transport Origin-Destination (O-D) matrices using smart card fare data. The main elementof connecting passenger’s trip-legs and infer O-D trips is a transfer time threshold. Different values of the transfer time threshold have been used in the past. The unique smart card faredata from South East Queensland, Australia, gives the opportunity to assess the transfertime and distance in a public transport network. This paper empirically investigates the use of South East Queensland (SEQ) data to study the effect of different transfer time thresholdson the estimated O-D matrices and provides a sensitivity analysis for different values. As a result, there is a slight effect on the inferred O-D matrices and transfer trips. Most of the transfer trips have one transfer where few transfer trips have two and three transfers. Morethan 90% of passengers walk less than 400m to transfer between alighting and the next boarding stop. The average transfer walking distance is between 45m and 200m where most of the transfers occur for the trips starting from the CBD area.

Keywords:

Transfer time threshold , O-D estimation , Inter-transaction time , Public transport smart card data