This paper describes a new method developed within the BEEP project (Big data for Environmental and occupational EPidemiology) to estimate road traffic flows and to improve the truthfulness of noise maps for agglomerations through Big Data treatment. This new approach, based on data provided by Google API, acquires information regarding travel time to estimate traffic volumes using link delay functions. To achieve this goal, an appropriate experimental plan was designed to simultaneously collect travel times by Google Application Programming Interface (API) and traffic volumes on site. The experimental survey, carried out in the cities of Rome and Pisa, involved different types of road links with traffic lights or roundabouts and different number of lanes. The influence of link characteristics on the correlation between travel time and traffic flow was analysed. The method developed was used in a small area of the city of Rome, and noise maps derived by Big Data were compared to noise maps produced via conventional means.