NLP Approach to Uncovering Consumer Sentiment and Topics on Biodegradable Sponges : Implications for Product Innovation and Sustainable Marketing

This study aims to analyze consumer sentiment towards biodegradable sponge products as an eco-friendly alternative to conventional sponges that contribute to the global plastic waste problem. Product review data were collected from several stores in the world's largest e-commerce consisting of more than 1000 random reviews. Using sentiment analysis and topic modeling based on Natural Language Processing (NLP), this study analyzed customer reviews of biodegradable sponges to identify consumer reactions and key topics discussed. The results of the analysis showed that most reviews tended to be negative (68.7%), with only 31.2% positive reviews and 0.1% neutral reviews. Monthly sentiment trend analysis showed an increase in the number of biodegradable sponge users since 2018, although negative reviews continued to increase each year. Latent Dirichlet Allocation (LDA) topic modeling identified two main topics in positive reviews, namely product effectiveness and customer satisfaction, while negative reviews discussed more about product size, durability, and material quality. Based on these findings, it is recommended to improve product quality by paying attention to the thickness, size, and durability of the sponge. In addition, marketing strategies that focus on product effectiveness and building relationships with customers can also increase sales. This research is expected to contribute to the development of environmentally friendly products and support the transition to more sustainable consumption.