Weight Management Short Video Quality Scores from TikTok and Bilibili

A cross-sectional study of 200 weight management-related short videos collected from TikTok and Bilibili. The dataset includes quality and reliability scores (GQS and DISCERN) assigned by researchers, along with video characteristics like duration, likes, and uploader type. It was created by Jinsong Du and last updated in June 2026.

Use Cases

• Training models to predict video quality based on features like duration and uploader type mentioned in the description.
• Analyzing correlations between video popularity metrics (e.g., likes) and professional reliability scores.
• Comparing content quality and source distribution between different short-video platforms (TikTok vs. Bilibili).
• Interpreting machine learning model decisions for quality prediction using the SHAP framework described in the study.

Strengths

• Includes 200 videos, providing a basis for comparative analysis between two major platforms.
• Features researcher-assigned quality scores (GQS and DISCERN) for systematic evaluation.
• Model performance is quantified with a reported AUC of 0.8072 for the XGBoost-based prediction tool.

Limitations

• Row count for the underlying evaluation data is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is small in scale at 367.2 KB, representing a limited sample of content.

Provenance

Source

figshare, authored by Jinsong Du.

Collection Method

Cross-sectional study collecting the top 100 relevance-ranked videos from TikTok and Bilibili, evaluated by two independent researchers.

Freshness

Last updated 2026-06-02 05:40:31; freshness should be verified.