Quantifying the Effect of Fuel and Traffic Regulations on Air Pollution : A Data-Driven Modelling Approach

Air pollution is a major global issue, and Quito, Ecuador the world's highest capital city faces significant challenges in maintaining clean air. This study examines how four key environmental policies introduced between 2005 and 2012 influenced air quality in both urban and semi-urban areas. The policies analyzed include: the switch to low-sulfur diesel in Quito in 2005, the nationwide rollout of cleaner gasoline in 2009, the Pico y Placa, the policy restricts vehicles from circulating during peak hours based on the last digit of their license plate numbers in 2010, and the adoption of cleaner diesel across the country in 2012.To assess the impact, we used air quality data from monitoring stations in Carapungo, El Camal, Belisario, Tumbaco, and Cotocollao. They applied machine learning techniques, such as Gradient Boosting, while accounting for weather factors like temperature, humidity, and wind speed. By creating weather-adjusted models, they could isolate the direct effects of each policy by comparing predicted pollution levels with real-world measurements during critical policy periods.The findings reveal that the cleaner diesel policies in 2005 and 2012 significantly reduced sulfur dioxide (SO₂) levels across multiple locations. However, PM2.5 reductions varied depending on local pollution sources and enforcement. The 2009 cleaner gasoline initiative led to slight improvements, particularly in high-traffic urban zones. Meanwhile, the Pico y Placa traffic restrictions in 2010 brought localized air quality benefits. Overall, the study demonstrates how data-driven methods and machine learning can effectively evaluate environmental policies and support efforts to achieve cleaner air in cities.