Application of Statistical Cluster Analysis in Design Process – the Case of Fabric Design

Abstract

There are many bed linen manufacturers in Taiwan. They want to success in the domestic bed linen collection market. However, the fabric designer always has the ideas with color and pattern by self-experience. For the fabric designer and manufacturer, they need a reference frame to communication. For this reason, we attempt to apply the statistical techniques for them to get more design understandings in the market of bed linen collections. We use the cluster analysis in a potential consumer survey to evaluate whether is the consumer’s lovely color and pattern with emotional preference.For the main bed linen buyers are the female in Taiwan market. The research focuses on 20-40 years old female interviewees. It collects more than 600 respondents’ data from three major cities in the northern, central and southern regions of Taiwan. The cities are Taipei, Taichung, and Kaohsiung. The interviewees choose the top 1 to 3 that they like the most and willing to buy from provided 16 photos of bed linen collection in different styles. Then, the result of the statistical cluster analysis shows that 16 collections can be grouped as 7 clusters with different favors. From the emotional preferences of color and pattern, we find out which collections are more suitable for different regions, and which ones are less popular among consumers. This can provide effectively the fabric designer with the direction of domestic consumer demand. And, also let manufacturer predetermine what combinations are suitable for some regional sales.In the past, fabric designers often used self-experience to design product. Now, we can try to conduct such surveys first, and then use more objective statistical results to help designers understand the bed linen design needs of consumers in different regions. At the same time, manufacturers can also better understand what types of products are suitable for sale in which region, also reduce inventory and slow-selling problems. Final, we hope to verify this statistical evaluation technique based on more manufacturers' sales data in the future, and whether it can improve their profits in practice.

Keywords: Fabric Designer, Bed Linen, Cluster Analysis, Evaluation

DOI: 10.54941/ahfe1005158