Improving food safety through data pattern discovery in a sensor-based monitoring system

Increased digitalization has enabled firms to better monitor and improve operations to meet the increasing demands of food safety. Sensor-based monitoring systems are widely adopted by food producers to improve quality management and facility monitoring capabilities. However, existing studies on data mining in food safety applications are typically focussed on the development of standalone models; thereby tasking users with high levels of manual data input or requiring extensive, customer-specific post model implementation endeavour. To address the existing gap, this paper proposes a transferable association rule model that enables scalability with actionable information and data insight. The model was developed using the CRISP-DM framework with datasets of more than 4000 alarm entries from a global fast food franchise. The proposed integrated system is capable of providing reliable visual information for multiple managers to monitor their respective quality performance. This study contributes to the existing body of knowledge by providing empirical insights in the context of digitalisation and food safety.