Research Projects

India is home to some of the world’s most polluted cities, with 9 out of the top 10, 12 out of the top 15, and 17 out of the top 20, according to IQAir. The sources of this pollution can be divided into perennial contributors, such as transportation, construction, and industrial emissions, and seasonal contributors like stubble burning, fireworks, and specific weather patterns, which worsen pollution during certain times of the year, especially in winter. For effective policymaking, it is critical to understand the main sources of pollution and their individual contributions. However, no comprehensive study has been done, either in India or globally, to accurately quantify how much each factor contributes to overall pollution. This lack of data makes it challenging to design targeted and effective interventions to address air pollution. Our project aims to tackle this problem by developing techniques to disaggregate air pollution, meaning to estimate the contributions of various sources to overall air quality. We will use publicly available air quality index (AQI) data from cities across India, sourced from platforms like Data.gov.in and the Central Pollution Control Board (CPCB). While this data provides an overall measure of air pollution, it does not indicate which specific factors are responsible. To bridge this gap, we will develop machine learning models to analyze AQI data and estimate the contributions of different factors, such as vehicular traffic, industrial emissions, and stubble burning. By identifying patterns in the data and correlating them with known pollution events, we aim to break down the overall AQI into its contributing components, helping to pinpoint which factors are most responsible for poor air quality in specific regions. The goal of this research is to help policymakers make informed decisions. For instance, if our models reveal that vehicular traffic is the largest contributor in a city, authorities can prioritize reducing vehicle emissions, perhaps through promoting electric vehicles or improving public transportation. If stubble burning is found to be the main cause of pollution in Northern India during the winter, policies could focus on promoting alternative agricultural practices or better stubble management. In conclusion, our work aims to fill a critical gap in current environmental research by developing machine learning techniques to disaggregate pollution sources. This will provide government officials with the data they need to create targeted interventions, helping improve air quality and public health across India’s most polluted areas.