Identifying Customer Preference Factors in Front-Warehouse Fresh E-commerce: A BERTopic and Sentiment Analysis Approach
Year: 2025
Type: article
Abstract: In recent years, front-warehouse fresh e-commerce has rapidly expanded in China, with platforms such as Dingdong Maicai meeting consumer demand for high-frequency and immediate grocery purchases through instant delivery.However, the perishability of goods, reliance on cold-chain logistics, and price sensitivity create divergent consumer experiences.Drawing on 14,118 user reviews segmented into 25,297 sentence-level units, this study applies BERTopic to extract themes and employs a RoBERTa-based sentiment classification model to identify polarity.A total of 38 valid topics were identified and consolidated into seven aspects: freshness, taste, packaging, delivery, price, customer service, and image-text mismatch.Results indicate that freshness and packaging are the primary sources of negative sentiment, while price and delivery attract both positive and negative attention.Overall, consumer evaluations are predominantly positive, yet product deterioration, damaged packaging, and inconsistencies between product descriptions and reality elicit notable dissatisfaction.This study not only provides data-driven evidence for understanding consumer preferences but also offers practical implications for platforms to optimize product management and operational decision-making.
Authors: Xiaomei Han, Hamrila Abdul Latif, Chin Hong Puah
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4414240223
FWCI:
Cites:
Cited by:
Related to: