Complex systems often involve many interacting state variables whose nonlinear dynamics can be chaotic. Modeling such systems is challenging because we typically observe only a subset of the state variables and often lack first-principles equations that govern the dynamics of these variables. This motivates data-driven approaches such as Empirical Dynamic Modeling (EDM), which uses Takens’ theorem to reconstruct the underlying dynamics from observed data. In this work, I apply EDM to model the dynamics of a large-scale power grid, focusing on three key variables: load (net demand), Area Control Error (ACE) (the mismatch between scheduled and actual power flows), and frequency (the grid’s response to imbalance). Although nominally 60 Hz, the grid frequency fluctuates within a narrow band due to continual mismatches between supply and demand.