Digital Phenotyping of Affect and Stress in Emerging Adults Over One Year

A one-year study of 24 emerging adults (ages 18–21) collected daily ecological momentary assessments and continuous smartphone sensor data. The dataset, authored by Coralie S. Phanord and shared under CC-BY-4.0, compares idiographic and nomothetic XGBoost models for predicting daily affect and stress from behavioral features like sleep, activity, mobility, and phone use.

Use Cases

• Comparing personalized versus general machine learning models based on smartphone-sensed behavioral features
• Predicting daily affect and stress levels based on passively collected sleep, activity, and mobility patterns
• Evaluating the stability and accuracy of idiographic modeling approaches for digital phenotyping
• Investigating the role of stable individual differences in predicting mood from smartphone data

Strengths

• Includes a one-year longitudinal data collection period
• Compares three distinct modeling approaches (Group General, Group Personalized, Within-Person Personalized)
• Uses daily ecological momentary assessment for ground truth labeling

Limitations

• Small participant sample size (n=24)
• Row count and column-level documentation are unknown
• Description metadata is limited; actual data quality requires manual inspection after download

Provenance

Source

figshare

Collection Method

Continuous collection of smartphone sensor data paired with daily ecological momentary assessment.

Time Range

One year

Freshness

Last updated 2026-06-04 05:41:53