Abstract:

With the growth of e-commerce, consumer reviews are becoming more widely available and influential. These valuable online product reviews (OPR) show any product issues and contain unique and hidden user information fragments that designers can use in decision-making. OPRs are often unstructured, massive, disorganized, and highly detailed. OPRs are voluntary production and are available in large numbers, publicly available, and accessible. These features increase the number of samples and save money and time for designers to understand the user. The analysis of OPRs is done with AI-supported text analysis tools, especially if many reviews are to get through. In this study, user demographics and opinions about the product are extracted through text mining and statistical methods through the OPRs of a sample product. The data analysis results provided valuable information about the users and had the potential to develop new knowledge and generate new ideas for the design process. By arguing for the merit of adding Big Data analysis to the design process, first, valuable user information content contained in OPRs has been revealed. Secondly, it was possible to express user stacks as clusters with similar characteristics. Finally, it has been revealed that demographic user clusters become homogenized after the product experience, and the initially disjointed clusters begin to resemble independently from the demographic clusters due to independent product/aspect evaluations.

Keywords: