We face a quickly changing world, where critical conservation decisions must be made based on limited ecological data. To make informed decisions, it is vital to understand the dynamics of ecological communities, and the underlying network of interactions that shapes those dynamics. As computing power continues to increase, we may benefit from a variety of sophisticated computational techniques, drawn from across the sciences, and use them to improve our ecological understanding. In the following studies, I use computational approaches to characterize the structure and dynamics of ecological communities, with a focus on the Tatoosh Island middle intertidal. The Tatoosh Island intertidal is one of the longest-studied systems in ecology; first studied by Robert Paine in the early 1960s, the system has been used to study the influence of predators, disturbance, and indirect effects in ecological communities. This is an excellent system for the study of network structure and dynamics, both because of the diverse community of organisms, and because of the rich data available, including a network with trophic and nontrophic interactions and a long-term dataset of community composition under control and experimental conditions. I use data from this and other communities, in conjunction with machine learning and other computational methods, to make inferences about the structure and dynamics of ecological communities.