Abstract
This research introduces an innovative module for air quality assessment that integrates machine learning models within a unified information and analytical platform. The proposed methodology combines air pollution data with chemical analyses of snow cover samples, specifically collected in Ust-Kamenogorsk, Kazakhstan. As a complementary bioindicator approach, watercress plants were cultivated using meltwater from these samples to observe biological responses to pollutant exposure. By correlating the chemical composition of airborne contaminants with the physiological reactions of the plants, the study offers valuable insights into the broader ecological impacts of air pollution. A variety of machine learning algorithms were applied to analyze the collected data, with model performance evaluated to identify the most accurate predictive approach. The findings reveal key pollutants that most significantly affect air quality and pose risks to public health in the region. As a potential direction for future work, the integration of augmented reality (AR) technologies is suggested to enhance the visualization of environmental data. An AR-based approach could offer interactive, spatial representations of pollution levels, plant-based bioindicators, and regional risk profiles, enabling users to intuitively explore and understand complex environmental patterns. This would improve public awareness, support decision-making, and further promote real-time air quality monitoring and communication strategies in Kazakhstan.