Modelling of traffic noise

Traffic noise is becoming a serious problem, not only because it harms human health but also because it hinders the environmental sustainability of cities. Most studies that analyze how traffic noise is related to human health depend on detailed noise data, but such information is often hard to get, especially in Eastern Europe. We address this lack of noise data by using machine learning techniques to estimate long-term road traffic noise levels obtained from 232 measurement locations across the five largest cities in Bulgaria. Our results show that noise levels vary considerably across urban areas and that, for example, the distance to major roads and railroads is a critical factor in traffic noise levels. We also found that showed that 95% of the population in the five cities experienced harmful noise levels. Our noise maps are freely available to the public and can be reused in other studies and for urban policymaking.