Obesity Risk Prediction for Chinese Children and Adolescents, 2017-2020

Chinese children and adolescents from 31 administrative regions were surveyed for physical activity, sedentary behavior, and sociodemographic factors. The dataset contains 38 candidate predictors used to develop and validate a machine learning model for obesity risk, with data from 35,016 participants in 2017-2018 and 3,495 in 2020. Author Zekai Chen published the supporting documentation under a CC-BY-4.0 license.

Use Cases

• Developing obesity risk calculators based on parental BMI and lifestyle factors.
• Analyzing the impact of moderate-to-vigorous physical activity (MVPA) on childhood obesity.
• Evaluating the predictive importance of screen time (mobile devices, TV) on weekdays.
• Comparing the performance of machine learning algorithms like Random Forest for health risk prediction.
• Creating large-scale screening tools for healthcare agencies and schools.

Strengths

• Data is based on a nationally representative sample of 35,016 participants.
• Model was validated temporally with a separate dataset of 3,495 participants.
• The study identifies and ranks the top 5 predictive features, including parental BMI and MVPA.
• A web-based risk calculator derived from the model has been deployed.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying data tables is unknown, which may limit suitability assessment.
• The dataset file is a 1.2 MB DOCX document; the actual structured data may require extraction.

Provenance

Source

Physical Activity and Fitness in China—The Youth Study (PAFCTYS)

Collection Method

Cross-sectional survey data collection.

Time Range

2017-2018 for development, 2020 for validation.

Freshness

Last updated 2026-05-05 04:12:51; freshness should be verified.

Geography

31 administrative regions across China.