Personal air pollution exposure: Monitoring methods and the emerging role of big data and machine learning

People are increasingly exposed to air pollution, which poses significant health risks, including respiratory and cardiovascular diseases. Research on air pollution exposure assessment has increased significantly since 2011, with notable contributions from China and the US. Air pollution exposure studies require high-resolution pollutant data to accurately capture the varying conditions individuals encounter in different environments. Nevertheless, currently conventional ambient air monitoring methods remain limited in their ability to capture the fine-scale spatial and temporal variability of air pollution. Based on a comprehensive evaluation of 246 representative studies, this review systematically assesses the current methodologies for personal air pollution exposure, with a focus on time-activity patterns, personal exposure monitoring (PEM), and GPS-based modeling. Although PEM provides high-resolution data closely aligned with individuals’ daily routines, it is limited by calibration complexity and environmental sensitivity. Time-activity pattern analysis and GPS-based modeling improve applicability and scalability yet rely on assumptions about human behavior and sensor or GPS signal limitations that may introduce uncertainty. To address these limitations, recent advances increasingly integrate multi-source data, Internet of Things (IoT), and machine learning technologies. These emerging approaches enable more dynamic, personalized, and equitable exposure assessments, enhancing both precision and reliability. This review not only highlights both the comparative strengths and weaknesses of existing methods and the promising opportunities of data-driven innovations to support targeted public health interventions and reduce exposure inequalities across populations, but also provides a consolidated reference to guide methodological improvements and future research directions in personal exposure studies. • Provide a systematic comparative evaluation of current methods for personal air pollution exposure assessment. • Clarify the roles and differences of time-activity patterns, personal monitoring, and GPS-based modeling in capturing spatial-temporal variability of exposure. • Future efforts should focus on multi-source data and machine learning to enable accurate, dynamic and personalized exposure assessment.