Data-driven analytical framework for waste-dumping behaviour analysis to facilitate policy regulations.

Waste sorting at the source is a vital strategy of waste management and to improve urban sustainability. If the strategy is implemented by relying solely on publicity and civic awareness, the impact is less significant. Proactive measures, such as policy regulations, supervisory guidance, and stimulating incentives, play essential roles for better management. The unknown waste-dumping behaviour of residents is a great challenge for decision-makers to allocate resources for waste-collection operations and to refine regulations. Traditional behaviour analysis methods such as questionnaire surveys and simulation methods have limitations considering the population size and the complexity of individual behaviour. This study aims to design a data-driven analytical framework to analyse household waste-dumping behaviour and facilitate policy regulations by using the Internet of Things (IoT) and data mining technologies. The analytical framework is further developed into a four-step management cycle. A case study in Shanghai is employed to demonstrate the effectiveness of the analytical framework and management cycle. The results of behaviour analyses reveal that (1) waste-dumping frequency is high in the evening but negligible in the early afternoon; (2) compared to working days, peak-value time at weekends occurs later in the morning and earlier in the evening; (3) residents require longer waste-dumping time windows than those empirically recommended by administrators. Managerial insights and decision support based on these research results have been presented for decision-makers to guide operations management and facilitate policy regulations.