The underlying spatiotemporal mechanism that leads to the formation of seizures in the brain has been an interesting topic for decades. Different techniques have been proposed to extract the dynamics of epileptic recordings that are involved in seizure formation. Methods have been used to measure the synchrony between two or more epileptic recordings. These techniques are often model-based or suffer from poor time-frequency resolution. In this project, we introduce a data-driven toolbox called the Dynamic Mode Decomposition (DMD) with time-delay embedding to extract the underlying spatio-temporal dynamics of seizure formation. These techniques will enable us to focus on similarities among seizures in our attempt to better understand, detect, and predict seizures. The inferred information on the underlying dynamics of an epileptic system are essential in terms of improving the capability of stimulation-based treatments of epileptic patients.