Application of text mining in PTT forum in analysis of consumer preference for online shopping platforms

ABSTRACT

With the advent of economic development and Internet technology, offline retail stores have gradually shifted to virtual shopping networks, and consumers' online shopping has become increasingly prosperous. Moreover, since the COVID-19 pandemic, the public has taken the initiative to reduce the number of outdoor activities, which has increased consumers' willingness to shop online. This research takes Shopee, PChome, and MOMO online platforms as the research subjects. We obtained data from 2020 to 2021 on PTT e-shopping and lifeismoney boards. In addition, we used web text crawling analysis, R data text mining and positive/negative sentiment analysis, and word cloud to determine popular keywords related to online shopping issues, and consumers' preferences for online shopping platforms are studied. The results show that "seller", "problem", and "offer" are the most discussed keywords indicating that people care about the consumer experience to a certain extent. The next most frequent keywords are "coat", "dress", "shopee", "discount", "cheap", "Taobao", and "Taiwan", which will appear according to the needs of consumers in different seasons. Based on the sentiment analysis, the consumers posted more positive articles than negatives in PChome (2.33) and MOMO (2.34) compared to Shopee (1.11). Through term frequency analysis, we can understand the trends and suggestions brought by popular keywords of online shopping to consumers and online store sellers, and also allow online store sellers to analyze the key decision concerns and the possibility of customers' behavior. © 2023, International Journal of Systematic Innovation. All rights reserved.