Ordinary graphs only support discrete structures. In this paper we present an approach towards embedding continuous data – like time stamps or series of measurements – in discrete graph models. These continuous meta-information implicitly define relations between vertices which are not explicitly defined in the graph itself. We call this an induced Non-Discrete Graph Structure (NoDeS). The model is helpful for visualization of time-dependent models or values from physical domains. We provide a formal definition of NoDeS based on graphs and two mappings, instance and annotation based, to already known graph structures and visualizations. A visualization of multi-partite projection provides a representation of information from several contexts, enabling NoDeS for a generic context switching mechanism which is used for interaction with these structures. Finally, we introduce an application concept for agent-driven event scheduling using NoDeS.