Dan Sheldon’s primary research develops algorithms to understand and support decisions about the environment using large data sets. He and his colleagues and students are the primary machine learning research leads for BirdCast. He seeks to answer foundational questions and also to build applications that transform large-scale data resources into scientific knowledge and policy. Some examples of his work include spatial optimization to conserve endangered species, continent-scale modeling of bird migration, and biological interpretation of weather radar data across the US. Methodologically, Dan’s primary interests are machine learning, probabilistic inference, and network modeling. His work has contributed broadly applicable new approaches for reasoning about aggregate data in probabilistic graphical models, and for optimization of diffusion processes in networks.