In this talk I will summaries some of our previous and current activities in big data visualization. I will talk about the visualization of space-time fields, physical quantities that are structured in space and evolve in time, with a focus on techniques that enable interactive visual exploration of such fields. Key to interactivity is the use of regular grid structures, in combination with level-of-detail representations, data compression, and GPU volume ray-casting. I will demonstrate the power of this combination in various applications, ranging from scientific visualization to previewing systems for visual effects. For fields that are not structured on regular grids but on scattered point sets or triangle meshes, I will discuss possibilities for restructuring.