This study aims to explore the utilization of user-generated content for product improvement and decision-making processes. In the era of big data, the channels through which enterprises obtain user feedback information are transitioning from traditional methods to online platforms. The original data for this study were obtained from customer reviews of cordless hairdryers on JD.com. The specific process is as follows: First, we used the Python Requests package to crawl 20,157 initial comments. Subsequently, the initial data were cleaned, resulting in 1405 valid comments. Next, the cleaned and valid comments were segmented into Chinese words using the HanLP package. Finally, the Latent Dirichlet Allocation (LDA) method was applied for topic modeling. The visualization of the topic clustering was generated using pyLDAvis, and three optimal topics were identified. These topics were named “User Experience”, “Product Evaluation”, and “Product Features”, respectively. Through data analysis and expert consultation, this study developed product design improvement strategies based on online reviews and verified the validity of the developed cordless hairdryer design index system through a questionnaire survey, providing practical references and innovative theoretical foundations for future product design assessments.

Universal design, as an important concept in product development, has been gradually implemented in various industries and firms, but most of the applications are aimed at the incremental innovation of product design. However, considering the differentiation of product types, the design factors are varied and not precise, which makes it difficult to effectively implement the existing universal-design principles in specific product designs when facing architecturally innovative product development, which leads to an increasing amount of time and resources. This study, with an umbrella stand as an example, proposes to design a new universal-design scale by combining the attributes of the umbrella stand, the existing universal-design principles and the usability principles at the beginning of the design. Then, through interviews and user surveys, cluster analysis is used to screen the products and refine the design factors. In addition, with the Quantification Type I, the universal-design principles and design factors are matched effectively in order to obtain the weight differences of the product design factors and eliminate unsuitable design factors, in order to guide product design. Finally, the universal umbrella-stand-design case is completed for verification. The results show that the universal umbrella stand has been significantly improved in the evaluation of universal-design principles. In the future design, architecturally innovative products with inaccurate design factors can also use this process to complete the design and development of the target product, and to significantly enhance the universal-design evaluation of the product.