Abstract: In order to effectively avoid the blind and disordered competition and product homogenization among enterprises in the industrial cluster, it is essential to improve the product layout design pattern and promote the transformation and upgrading of industrial clusters. To this end，a data-driven design method is proposed for the product layout of industrial clusters. The proposed method constructs a data-driven design method and decision supporting model for product layout of the industrial cluster, and depicts the distribution characteristics of cluster products from the following four dimensions, structure and performance parameters, innovation capabilities, product prices and space innovation density. The proposed method was based on the data resources of the industrial cluster, clustering algorithms and text mining techniques are used to provide decisions supporting for the layout design of industrial cluster products. In other words, based on four dimensions of product structure performance, product price, enterprise innovation ability and spatial innovation density, a series of techniques and methods related to product layout evaluation have been studied for industrial cluster and empirically studied the distribution characteristics of industrial cluster product layout. The research in this paper highlights the powerful supporting role of big data in industry and enterprise decision-making and design, which could help decision maker to analyze and observe the current situation of industrial clusters and enterprises from a more scientific and comprehensive perspective, and completely change the logic of product layout design of traditional enterprises, and provide a new thinking model about enterprise innovation. The framework of multi-dimensional decision model provided a comprehensive perspective for enterprise managers to study and judge the situation of industrial cluster or enterprises. The reference value of different dimensions for product layout decision was well illustrated by examples. The phenomenon and level of product homogeneity of cluster enterprises was revealed through the clustering analysis of product structural performance data and the combination of product price data, which proved the existence of this phenomenon through the example of Ningbo injection molding machine industry. From the view of industrial ecology, the reasonable regulation and balancing on product layout among industrial cluster member enterprises are necessary. In addition, through the spatial innovation density analysis, the radiating and leading capacity effect of leading enterprise on cluster innovation performance was effectively reveals. Through the example of injection molding machine, a very important conclusions could be drawn that the trend on technical innovation of injection molding machine cluster presents a phenomenon of more dispersed or rare gathered. While, for the innovation and development of industrial cluster, members of industrial cluster must improve the ability on innovation resources distribution and efficiency of the transformation of scientific and technological achievements. Obviously, innovation is still one of the most important driving forces for the development of industrial clusters. On the other hand, from the perspective of knowledge value spillover, the sustained support to leading enterprises should be constantly reinforced, and improve their leading role in innovation, the effect of leading enterprises on innovation is obvious and remarkable in the regional of industrial clusters highly developed. In addition, the research result of the paper could provide valuable reference to the regional industrial alliance, industry associations and government sectors that the reasonable guidance for the positioning of innovation output of cluster enterprises is of great significance to the benign development of the industry clusters. At the same time, it is meaningful to realize the complementary of advantages and reasonably arrange the core competitiveness at the cluster level. Furthermore, the research was deeply integrated with large data analysis technology. So, combined with the patent map, the research pattern was used to help enterprises to identify the innovation research hotspot of each enterprise, update the pain points of technology innovation and the blind spot, and transfer data resources into knowledge service, pushed the service to cluster enterprises, associations and government agencies, etc. On the whole, the research could provide scientific decision support for the overall arrangement of product differentiation competition advantage, and provide solution for the construction of a good cluster innovation environment.

Key words: data-driven, product layout design, industrial cluster