Data-Driven Framework for Validating Policies: Air Quality Case Study

Abstract

This paper provides a data-driven framework for validating policies in transportation systems using machine learning techniques by using air quality and clean air zone policy as a case study. The underlying idea of our framework is to compare the behaviour of the transport system under study in two cases of with and without the policy. Historical data and new data gathered after applying a policy can be used to construct machine learning models and then make the comparison. This comparison will then be judged against the quantitative objectives of the policy. We employ the dataset from the Urban Observatory in Newcastle, United Kingdom, and consider a policy related to clean air zone with the objective of reducing the concentration of NO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> . We develop a recurrent neural network model called long short term memory (LSTM) for predicting the behaviour of the NO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> after the implementation of the charging clean air zone. Our implementations show that the LSTM model can successfully predict the NO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> concentration with root mean square error of 0.95. These results show the potential of using data-driven methods in analysing and validating policies in transportation systems.