A multitude of dynamical systems are described by a set of a large number of nonlinear differential equations which poses significant challenges in model characterization and analysis. This is either because it is very expensive to solve these equations using nonlinear analytical models or due to the magnitude of the model errors when applied to real-world applications. Our group develops unique data driven and machine learning methods to tackle this problem and identify simplified models of the dynamics of complex systems using data sampled from system dynamics. The methods facilitate the analysis of both high-fidelity models and real-world systems.