Air pollution kills seven million people around the world every year through stroke, heart disease, lung cancer, and acute respiratory infections. The most dangerous form of air pollution is PM2.5, or particulate matter less than 2.5 microns in diameter, which is the sixth highest risk factor for death around the world. In New York City, more than 3,000 deaths are attributed to PM2.5 each year with low-income neighborhoods and communities of color most impacted. Since PM2.5 is highly localized, it is extremely important to consider the number and locations of air pollution sensors to properly inform street-level policy. By collaborating with local schools, community-based organizations, and research labs, I deployed a citizen science network of low-cost PurpleAir sensors across New York City and compared its performance to expensive existing City- and State-run monitor networks. I then used machine learning to determine the optimal locations for the PurpleAir sensors and propose recommendations for future data collection and air pollution policy in New York City.