Analytical Outputs from Cross-Platform Study of Chinese Inkstone Markets

3,652 user-generated texts from five Chinese digital platforms were analyzed for a study on intangible cultural heritage markets. The dataset contains aggregated analytical outputs including LDA topic assignments, LSTM sentiment classifications, and statistical test results. Raw text data is not included due to platform terms of service restrictions.

Use Cases

• Validate topic modeling techniques based on the provided LDA topic-word probability distributions and evaluation metrics.
• Benchmark sentiment analysis models using the included LSTM model performance metrics and sentiment classifications.
• Examine cross-platform discourse patterns using the cross-tabulation tables and chi-square test results.
• Replicate statistical analyses of consumer behavior with the provided ANOVA and chi-square test results.

Strengths

• Includes analytical outputs from 3,652 user-generated texts across five distinct platforms.
• Provides multiple validated model outputs, including LDA topics, LSTM sentiment, and statistical test results.
• Dataset is licensed under CC-BY-4.0, allowing for broad reuse and sharing.

Limitations

• Raw user-generated text data is not included, preventing direct analysis of the source material.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the analytical outputs is unknown, which may limit suitability assessment.

Provenance

Source

User-generated texts collected from Taobao, JD.com, Xiaohongshu, Xianyu, and Kongfuzi.

Collection Method

Texts were analyzed using LDA topic modeling and LSTM sentiment classification, with subsequent statistical testing.

Freshness

Last updated 2026-05-11 06:14:41; freshness should be verified.

Geography

China (focus on Chinese digital platforms and intangible cultural heritage markets).