This project aims at calibrating a low cost sensor to make accurate prediction on Ozone levels. Frequentist and Bayesian frameworks have been used. Data have been analyzed by looking at correlation among variables and analyzing how they change over the time of the day.