Source-receptor mapping is the process of determining which populations are affected by emissions from a set of pollution sources. In the case of power plants, their emissions chemically react in the atmosphere and are transported long distances, impacting the health of exposed populations over vast geographic areas. Knowledge of source-receptor mappings is essential for informing policies aimed at reducing the public health burden of air pollution. Despite the importance of these decisions and the availability of high-quality data, data science has been undervalued in regulatory decision making in favor of deterministic chemical transport models (CTMs) for the movement of air pollution through space. In this paper, we apply a data-driven approach to this problem, using data science and statistical methods to perform source-receptor mapping by using only observed data to link emissions from power plants to exposed populations. Specifically, we use daily sulfur dioxide emissions from 385 coal-fired power plants and assess whether there is a statistical association with the daily fine particulate matter at 732 air quality monitors in 2005.