Inferring Exhaust Gases Levels using Taxi Service and Meteorological Data: An Experiment in the City of Porto, Portugal

Abstract

© 2016 IEEE. In this work, we study exhaust gas concentration patterns in Porto, Portugal, and explore techniques to estimate the level of NO2 concentrations using taxi service and meteorological data as sensors. The exploratory analysis revealed daily and seasonal patterns of exhaust gases, with higher concentrations in the morning and on colder months. Based on nine months of data, we are able to estimate the concentration of NO2 using a multilayer perceptron (r = 0.7358) and linear regression (r = 0.5407). The analysis was extended to estimate NO and NOx, showing a lower performance.