Wave Energy Converters (WEC) have great potential to help meet global energy demands, but even a single device can be challenging to model. Although most WEC concepts are modelled, many have been simplified with assumptions that do not accurately capture the true dynamics of the system. For this reason, data-driven modelling may benefit WEC development. In order to investigate these data-driven methods, a scaled physical model of an Oscillating Water Column (OWC) WEC was designed, constructed, and tested at Oregon State University. Using the data from the physical tank tests, several system identification methods were used to capture the nonlinear dynamics and develop state-space models of the OWC. A built-in MATLAB state space estimator, Eigenvalue Realization Algorithm (ERA), and Dynamic Mode Decomposition (DMD) were applied to the data. When compared against the validation data set, the state space model performed best, followed closely by the DMD method. These data-driven modelling methods allowed for any nonlinear dynamics to be captured in the state. Continuing to improve data-driven models such as these can help improve prediction and support ongoing controls work for WECs.