Applying text mining to analyze POP MART's popular product trends and business recommendations
Ming-Kai Hsu, Yu-Jen Pan · IET conference proceedings. · 2025
This study investigates consumer behavior and store display strategies for POP MART in Taiwan, utilizing Google reviews of their physical stores. Employing text mining techniques including TF-IDF, LDA topic modeling, and co-occurrence network analysis on the reviews, the research identifies popular IPs, purchasing patterns, and provides recommendations for store displays. The findings reveal LABUBU and MOLLY as the most popular IPs. While CHAKA ranks third in review volume, this is likely due to exhibition events, suggesting exhibitions effectively boost less popular IP visibility. Consumers show strong interest in blind box draws but express dissatisfaction with long queues and scalpers, indicating the need for service and on-site management improvements. Co-occurrence analysis shows frequent co-mentioning of DIMOO and SKULLPANDA. Based on these insights, the study recommends prioritizing LABUBU and MOLLY in store displays, bundling DIMOO and SKULLPANDA for sales, optimizing store flow and checkout, and using exhibitions to promote less popular IPs to enhance operational efficiency.