Planning public transport highly relies on the availability, quantity and quality of travel demand data of passengers. In the last two decades, smart card data has provided the opportunity to create comprehensive travel demand data as a byproduct of a fare-collecting system. One important attribute for the planning is the purpose of the trips, which is missing from the smart card data. This research study proposes and formulates a novel method to infer trip purpose in smart card data. Previous methods either lack the concept of trip chains or did not consider both spatial and temporal perspectives of a trip. Firstly, this method discovers relations between the sequence and temporal attributes of trips with their trip purpose attribute by running a clustering method on a rich travel survey dataset (This study only uses public transit records.) that contains all attributes. Secondly, the discovered clusters are labelled and transferred to the smart card data by calculating the closeness of the trip chain of each individual in the smart card data to the clusters. Thirdly, the proportion of relevant land use types near the destination of each trip is utilized to enhance the previously calculated closeness. The proposed method is implemented on datasets from South East Queensland, Australia. Also, two recently published methods were replicated and run on the same datasets to evaluate the proposed method. The results show improvements in the proposed method compared to the existing methods of the literature
