Prediction of Environmental Parameters for Smart Cities

Smart Cities are increasingly oriented towards managing environmental aspects and assessing the degree of pollution by using sensors and their prediction/reconstruction in all part of the city. One of the main objectives, in this context, is to establish and predict the air quality levels, such indexes strongly depend on the production of pollution connected to the traffic congestion in the urban areas. Typically, the sensors one can deploy on the city are limited, for their cost of installation and maintenance. Tools for reconstructing pollution level in non-directly monitored points would be very important. In this paper we show a methodology to make predictions about the pollution produced by the traffic flow and the air quality indexes in a city using a set of data including a sensor network infrastructure. The results obtained are applied in the city of Florence in the context of the TRAFAIR CEF EC project, starting from the traffic flow reconstruction algorithm produced in the Sii-Mobility MIUR project and using the Snap4City Big Data Infrastructure.