Humanual-Book: Amazon Book Reviews for User Behavior Simulation

36,622 comments from 209 frequent Amazon customers on book content, reflecting user preferences and tastes. This dataset is part of the HumanLM benchmark for training user simulators that accurately reflect real user behavior. The data was sourced from Amazon Reviews 2023 and spans from 1998-01-25 to 2023-05 10.

Use Cases

• Train user simulator models based on expressed satisfaction or dissatisfaction.
• Analyze customer sentiment and preference patterns in book reviews.
• Benchmark natural language generation systems for realistic user feedback.
• Study long-term trends in user opinions from 1998 to 2023.

Strengths

• Contains 36,622 comments from 209 distinct users, providing multiple perspectives.
• Covers a long time range from 1998 to 2023, allowing for temporal analysis.
• Part of a defined benchmark (HumanLM) for user simulation, suggesting a curated purpose.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the primary table is unknown, which may limit suitability assessment.
• Data may reflect temporal and source bias inherent to Amazon reviews from frequent customers.

Provenance

Source

Amazon Reviews 2023

Collection Method

Likely scraped or aggregated from Amazon review platform.

Time Range

1998-01-25 to 2023-05-10

Freshness

Last updated 2026-02-13 04:25:20; freshness should be verified.

Geography

Region 'us' is indicated in platform tags, suggesting a focus on the United States.