In May 2018, the World Health Organization (WHO) released new estimates of global air quality showing that air pollution levels are dangerously high in many parts of the world. Major sources of air pollution include the inefficient use of energy by households, industry, the agriculture and transport sectors, and coal-fired power plants. In some regions, sand and desert dust, waste burning, and deforestation are additional sources of air pollution. The new estimates reveal an alarming toll of 7 million deaths every year can be associated with exposure to outdoor and household air pollution, and that 90% of people worldwide breathe polluted air. More than 4,300 cities in 108 countries are now included in WHOs ambient air quality database, making this the world·s most comprehensive database on ambient air pollution. However, although air pollution monitoring is increasing, there remain areas for which information isn·t available and estimates of exposures are required for all areas. In this presentation we will discuss how we have been working with the WHO to develop the Data Integration Models for Air Quality (DIMAQ). This combines information from a number of different sources to allow exposures to be estimated worldwide. By integrating measurements from ground monitoring with information from satellites, population esti- mates, land-use and other factors to allow us to provide estimates of air quality for every country and region, including those where there is little, or no, monitoring. We describe the progression of the development of a series of models from spatial to spatial-temporal and their implementation on a global scale. We will present the findings from the most current analysis of the state of global air quality, using the current version of DIMAQ, including an examination of global, regional and country-level exposures and health burdens. We see show that there is great variability in air pollution across the world, with some areas experiencing levels that are more than 5 times higher than the guidelines.