Popular usage of EZ-Link cards for transit and non-transit purposes generate large data sets that are rich in information and can be mined using big data analytics. The transaction data provide an opportunity for uncovering hidden patterns and correlations for insights.
The current project is a joint effort to develop two data analytics applications on EZ-Link transaction data for: (1) fraud detection through abnormal usage pattern discovery; and (2) customer persona development using account-based data.
Data scientists of IDS and domain experts of EZ-Link are collaborating to get to the bottom of fraud detection. Domain experts provide ad-hoc or potential rules for abnormal usage, and data scientists will enhance the human expertise with newer unforeseen rules on pattern usage extracted from the data. The data scientists will also develop algorithms to infer different customer persona’s POIs and usage behavioural patterns and extend them into urban-scale POIs and patterns.
IDS and EZ-Link had an earlier joint effort (Project “EZ-Transaction Analytics for a Cashless Future”, 2017-18) to understand EZ-Link customers’ top-up behaviours. The two applications from the current project will continue to harness EZ-Link’s transaction data and translate them into ready-to-implement solutions for EZ-Link.
Project Title: EZ-Link Customer Persona Development and Abnormal Usage Pattern Discovery
Done By: EZ-Link Pte Ltd, Institute of Data Science (IDS) and Singapore Data Science Consortium (SDSC)
Contact Person: Prof. Ng See Kiong, Deputy Director, IDS
This research is supported by the National Research Foundation, Prime Minister’s Office, Singapore under its Industry Alignment Fund (Pre-positioning) Funding Initiative.