Many real-world systems, such as social interactions, financial networks, and biological processes, can be represented as graphs. Graph Neural Networks (GNNs) exploit this structure by propagating information along edges, allowing node and graph representations to be updated through message passing. In this way, GNNs capture both local interactions and global patterns. My research applies this framework to uncover hidden connections and coordinated dynamics in complex systems, enabling the detection of subtle dependencies, the prediction of emergent behaviors, and a deeper theoretical understanding of connectivity.