Real-Time Event-driven Air Quality Inspection Framework for City-wide Pollution Level Monitoring

The World Health Organization (WHO) reported in 2019 that approximately 4.2 million premature deaths globally were due to ambient air pollution. A majority of these occurred in low to middle-income countries. This research project focuses on the development of a reusable real-time Event-driven Air Quality Inspection (EAQI) framework, to assist in the development of distributed sensor systems for collecting and analyzing pollution sensor measurements. The design is based around event streams, which store immutable raw data as events, providing a history of entities as well as latest values. This framework is designed for scalability, customizability, and ease of integration with data analysis methods. While the main objective of this project was to provide an application framework to facilitate public accessibility to air quality information, the framework design serves as a reusable template for developing new applications that require the scalability and inherently data-centric nature of this design. The framework was tested by building a representative, web-based application, which was tested to assess its fault tolerance, data validation and responsiveness. A user experience survey assessed characteristics of this application and users' views on its feasibility. The results were generally favorable, such as 69% of data request response being within 800 ms. In testing of its robustness, the application started to experience request timeouts beyond 500 concurrent users. Further work includes design additions for meteorological data feeds and plugin components.