Shopee Vietnamese Product Reviews with Sentiment Labels

Shopee Vietnamese Product Reviews Sentiment contains approximately 10,000 Vietnamese-language product reviews from the Shopee platform. The reviews are labeled based on their semantic content rather than user-provided star ratings, providing a distinct sentiment signal. The dataset is suitable for beginners in data analytics and deep learning tasks.

Use Cases

• Train a sentiment classifier on Vietnamese review text using the provided semantic content labels.
• Analyze the relationship between semantic sentiment labels and other potential review features for e-commerce insights.
• Fine-tune a language model for Vietnamese text understanding on a labeled corpus of 10,000 Shopee reviews.

Strengths

• Contains approximately 10,000 labeled data points for analysis.
• Labels are derived from semantic content, offering an alternative to potentially biased user ratings.
• Specifically focuses on Vietnamese language text, a valuable resource for NLP in that language.

Limitations

• The dataset size of ~10k rows is relatively small for training complex deep learning models from scratch.
• Specific column structure and available features beyond review text and label are unknown.
• Potential geographic bias as all data originates from the Shopee platform in a specific region.

Provenance

Source

Kaggle

Collection Method

Scraped or collected from the Shopee e-commerce platform.

Geography

Vietnam