Douban Movie Reviews for Sentiment Analysis

Douban Movie Review dataset contains user-generated movie reviews from the Chinese social media platform Douban. The dataset, created by 'dirtycomputer' and hosted on Hugging Face, was last updated in February 2023. It is categorized as having 10M to 100M rows and is intended for text and tabular analysis.

Use Cases

• Train sentiment classifiers using review text and associated sentiment labels.
• Analyze user opinion trends by linking review text with metadata like movie titles or user ratings.
• Perform cross-lingual sentiment analysis by comparing features in Chinese review text with other language datasets.
• Study the relationship between review text length and sentiment polarity scores.

Strengths

• Dataset size is categorized between 10 million and 100 million rows, indicating a substantial volume of data.
• Includes both text (reviews) and tabular (metadata) modalities for multi-faceted analysis.

Limitations

• Specific row count, column names, and data schema are unknown, complicating direct usage.
• Potential geographic and language bias as data originates from a Chinese platform (Region:us tag may be incorrect).
• Data freshness is unclear as the last platform update was over a year ago.

Provenance

Source

User-generated content from the Douban social media platform.

Collection Method

Scraped or collected from the Douban website, method details unspecified.

Time Range

null

Freshness

Last updated on Hugging Face on 2023-02 16; update frequency is null.

Geography

Primarily China, based on the source platform and 'Chinese Language' tag.