Current techniques for process synthesis aim at creating customized process plants in which production capacity is specified to meet certain market demand. However, it is known that, in most cases, significant economic savings can be achieved if standardized module-based designs can be developed instead of customized designs. In this paper, a novel framework is presented to generate a set of modular designs that are optimal for different ranges of market data based on customer requirements through integration of data-analysis and design/synthesis-optimization stages that have been traditionally performed separately. The basic idea is to apply a clustering methodology and design/synthesis optimization iteratively, allowing repartitioning of data based on design feasibility and a new optimization search based on the current clustering of data. The proposed approach expands the boundaries of design optimization to incorporate demand data analysis reflecting customer requirements. A detailed case study of a cryogenic air separation plant with current demand data is presented to illustrate the importance and practical relevance of the proposed approach.