Fooreview Product Sentiment Dataset

Fooreview contains over 100,000 product reviews from US consumers for sentiment analysis and text classification. The dataset was created by sangmini and last updated in December 2021.

Use Cases

• Train a text classifier to predict review sentiment labels using the review text column.
• Analyze consumer behavior patterns by correlating review text with product metadata.
• Fine-tune a language model on the review text for domain-specific NLP tasks.
• Perform feature engineering on tabular review data to identify key sentiment drivers.

Strengths

• Over 100,000 rows of review data, enabling model training.
• Includes both text and tabular modalities for multi-modal analysis.

Limitations

• Data is over two years old and may not reflect current consumer trends.
• Specific column names and label distributions are unknown, complicating direct use.

Provenance

Source

Hugging Face user sangmini.

Collection Method

Collection method unknown; likely aggregated from US-based review sources.

Time Range

null

Freshness

Last updated December 2021; static dataset.

Geography

United States (inferred from 'Region:us' tag).