Digitalization is increasingly critical to modernizing Swedish municipalities and enhancing societal services. The Stockholm municipality recently implemented a forecasting system to assess air pollution and associated health risks. This system, available via a mobile app and online since 2022, provides an Air Quality Health Index (AQHI) that reflects cumulative exposure to NOx, O3, PM10, and birch pollen, helping residents make informed choices about outdoor activities. Currently, the Swedish Transport Administration leverages this system to forecast pollution impacts on roadside air quality and to support environment-sensitive traffic management, particularly along the E4S corridor. This project will extend iHorse, a social innovation initiative (2022-2024) focused on advancing air pollution and health risk forecasts through data-driven models. iHorse introduced machine learning models to enhance traditional deterministic forecasts for mid- to long-term air quality predictions. The goal of this project is to develop a high-resolution, long-range forecasting system by integrating traffic and IoT sensor data with advanced machine learning models, including transformer and graph neural networks. This system aims to address key limitations in the existing AQHI predictions, particularly regarding spatial resolution. Presently, AQHI relies on data from stationary measurement points, which do not sufficiently capture urban spatial variability in pollutant concentrations. However, with the recent proliferation of IoT sensors in Stockholm, air quality data is becoming increasingly granular. The proposed system will leverage real-time traffic data from various detectors and high-density IoT air quality sensors deployed across Stockholm. By enhancing the iHorse model with spatial correlation factors, we aim to capture traffic patterns as a primary pollution source and improve prediction accuracy. This approach will allow for precise, spatially representative air quality forecasts that better inform public health and traffic management decisions.