<p>Dynamic networks, which capture the evolving interactions among entities, have flourished in various scientific fields. We propose a framework for exploratory analysis of these dynamic networks by representing them as latent functions. This framework comprises several visualization tools based on functional data analysis, specifically tailored for addressing typical tasks such as community detection, central node identification, and change point discovery. Besides, we develop a computationally efficient algorithm to obtain the latent functions. Through comprehensive simulation studies conducted under commonly investigated settings, we demonstrate the effectiveness of these tools. Furthermore, we apply the proposed tools to three representative and intriguing real-world networks, yielding enlightening discoveries. An R package for implementing the proposed methods, along with supplementary materials for this article, is available online.</p>