Analysis of Yearly Characteristics of Chinese Online Text Data on K-Beauty Using Text Mining: Focusing on 2021 to 2025

This study analyzes yearly characteristics of Chinese online discourse on K-Beauty from 2021 to 2025 using text mining. A total of 21,006 valid documents were collected from Baidu, Weibo, CNKI, and Google via TEXTOM. After preprocessing with Jieba segmentation, noun and adjective extraction, synonym unification, and stopword removal, the top 50 keywords were identified each year. Analyses included word frequency, TF-IDF, semantic network analysis, and CONCOR clustering. Results show that core keywords such as “Korea”, “cosmetics”, “Korean cosmetics”, “beauty”, “good”, “video”, and “Weibo” consistently appeared with high frequency and TF-IDF values across all five years, indicating that K-Beauty discourse is persistently centered on Korea-related terms and positive evaluative language, and is disseminated mainly through video-oriented social media. Yearly shifts are also evident: 2021–2022 feature multidimensional frames combining industry ecosystem, media experience, and product evaluation; 2023 emphasizes China’s market, price competitiveness, and domestic brands; 2024 highlights the technology-related keyword “microneedle” and pop-culture icons; and 2025 focuses on global markets, e-commerce, consumer experience, and industry performance. The study expands K-Beauty research from survey-based approaches to big-data text analysis and offers practical implications for branding, pricing, and digital marketing strategies in the Chinese market. Limitations include the failure to systematically exclude promotional posts, reliance on a single text-mining tool, and the absence of platform-specific comparisons and sentiment analysis.