Benjamin Laufer ’19
Operations Research and Financial Engineering
Using Machine Learning to Ensure Clean Air Act Compliance: Optimizing EPA’s Inspection Strategy
I worked with the EPA’s data-management team on using statistics and machine-learning methods to help the air-compliance branch with inspection targets. Currently, EPA inspectors use personal discretion in deciding which facilities to check for illegal air pollution-emission practices. Using an array of historical EPA compliance data, I aimed to devise a data-driven, optimal strategy for Clean Air Act inspections. With R programming, I implemented a Random Forest Machine Learning algorithm to classify suspect facilities nationwide. I hope my work will help the EPA identify and regulate sites that pose significant environmental risk to American citizens. My summer internship experience taught me a lot about American environmental regulations, and helped me hone my data-analytics and programming skills. Back at Princeton, I am continuing coursework related to mathematical modeling, big data and environmental policy. Additionally, I am continuing this project under the guidance of Robert Vanderbei, Professor of Operations Research and Financial Engineering.
U.S. Environmental Protection Agency, New York
Daniel Teitelbaum, Environmental Policy and Data Analyst, EPA