Multidimensional Network Analysis for Analyzing and Predicting Complex Customer-Product Relations in Engineering Design

Project summary Overview The objective in this research is to develop a multidimensional network analysis approach rooted in social network analysis for analyzing and predicting complex customer-product relations in supporting engineering design decisions. As shown in Fig.1 using vehicles as an example, customer-product interactions form a complex socio-technical system, not only because there are complex relations between the customers (e.g., social network) and between the products (e.g. product family or market segmentation), there also exist multiple kinds of relations between customers and products (e.g., “consideration” versus “purchase”). The premise of this research is that, similar to other complex systems that exhibit dynamic, uncertain, and emerging behavior, customer-product relations should be viewed as a complex socio-technical system and analyzed using social network theory and techniques. The structure and topological characteristics identified from consumer-product networks can discover emerging patterns of consumer-product relations by taking into account the heterogeneities among customers and products. Intellectual Merit Analytical modeling of customer preferences in product design is inherently difficult as it faces challenges in modeling heterogeneous human behavior. The proposed research will be one of the pioneering works in extending social network analysis for analyzing and predicting customer preferences in product design using complex, large-scale, and dynamic customer-product relational data. The novelty of the research lies in the employment of a multi-layer multidimensional network framework, where separate networks of “customers” and “products” are modeled in multiple layers, and multiple classes of between-layer customer-product relations are considered. We will demonstrate that a wide range of network structures can be used to analyze product associations and similarities, study the effects of product features and customer attributes in forming network structures, and identify new opportunities for design improvements. By extending the exponential random graph models (ERGM) from analyzing unimodal network to analyzing bipartite and multidimensional networks, our research will enrich ERGM as a unified framework for analyzing customer-product relations as well as predicting unknown customer preferences (consideration or choice in this work). A multidimensional network approach will be developed to measure simultaneously within-layer customer relations together with between-layer customer-product relations for assessing the social impact on customer choice behavior. This is especially useful for modeling marketing penetration of engineering products such as the adoption of cutting-edge technology products and sustainable products. By exploring the use of text analysis of customer-generated data via Web 2.0 and verifying the consistency with results from survey data, our research will create crowdsourced “virtual labs” for advancing computational social science in product design. Broader Impact Our research follows the basic principles of social and economic theory and is expected to extend the extant knowledge in engineering design as well as in market research, social network analysis, and large-scale data analysis. First, the results of our study have direct impacts on understanding complex customer-product relations in engineering design. It will help industry produce more competitive products in shorter time to market, considering not only the engineering requirements but also the heteroge