Scientific data processing and visualization of computational simulation data have played an important role in knowledge creation, and its transfer for the benefit of society, through scientific discovery and understanding of physical and chemical phenomena. Massively parallel processing architecture has greatly contributed to increase the computational power and to enlarge the scale of computation. This facilitated the generation of larger and larger amounts of data. To tackle this oncoming tsunami of data, we propose a post-processing system for large-scale data focusing on the idea that a postprocessing system needs to be an assistant of the researchers thinking process. We present a probable design solution targeting continuous development, and the technologies for productivity improvement such as collaborative work and visualization. This paper describes the current development status of a postprocessing system designed for the next-generation 10-petaflop supercomputer under development in Japan.