Abstract: Researchers are often interested in drawing causal conclusions from data. In many modern applications, data are structured over space, time, or networks, and may be statistically and causally dependent. Such dependence poses challenges for standard causal inference methods, but also creates new opportunities. In this talk, I will present an overview of my research on causal inference with dependent data. First, I show how structured data can be leveraged to relax the classical assumption of no unmeasured confounding. I then discuss methods for causal inference under interference, where the treatment of one unit may affect the outcomes of others, and illustrate how such effects can be estimated in dependent settings. I introduce a general causal inference framework for spatio-temporal point pattern data. Throughout the talk, I emphasize unifying principles and practical implications, highlighting how causal questions can be addressed in settings where dependence is present.