The consistent and accurate monitoring and modeling of urban ambient air quality are essential to reduce risks to public health. The growing number of open data sources now allow for data assimilation algorithms to improve air quality monitoring, including uncertainty estimates. The robust assimilation of model data with observations for urban ambient air quality monitoring requires the estimation of uncertainty to cope with unknown events and changing environmental conditions. However, uncertainty estimates from open access model simulations and observations are frequently unavailable. To address this gap, we propose a lightweight framework for ambient air quality data assimilation suitable for low-powered embedded hardware used by the Internet of Things. The proposed framework includes computationally efficient assimilation algorithms for the data-driven estimation of unknown uncertainty. The algorithms are compared and validated to assimilate open-access urban ambient air pollution observations and global numerical model simulations from the System for Integrated modeLling of Atmospheric composition (SILAM). This work is significant because it offers a computationally lightweight approach to sequentially assimilate open time series data without prior uncertainty estimates.