Group Health Dataset (Sleep and Screen Time)

Group Health (Sleep and Screen Time) Dataset

Title: Group Health (Sleep and Screen Time) Dataset

Description: This dataset includes biometric and self-reported sleep-related information from users wearing health monitoring devices. It tracks heart rate data, screen time, and sleep quality ratings, intended for health analytics, sleep research, or machine learning applications.  

Creator: Eindhoven University of Technology

Version: 1.0  

License: CC-BY 4.0  

Keywords: sleep health, wearable data, heart rate, screen time, sleep rating, health analytics  

Format: CSV (.csv)  

Size: 301,556 records

PID: 10.5281/zenodo.15171250 

Column Descriptions

- Uid (int64): Unique identifier for the user. Example: `2`

- Sid (object): Session ID representing device/session (e.g., wearable device). Example: `huami.32093/11110030`

- Key (object): The type of health metric (e.g., 'heart_rate'). Example: `heart_rate`

- Time (int64): Unix timestamp of when the measurement was taken. Example: `1743911820`

- Value (object): JSON object containing measurement details (e.g., heart rate BPM). Example: `{"time":1743911820,"bpm":64}`

- UpdateTime (float64): Timestamp when the record was last updated. Example: `1743911982.0`

- screentime (object): Reported or detected screen time during sleep period. Example: `0 days 08:25:00`

- expected_sleep (object): Expected sleep time duration (possibly self-reported or algorithmic). Example: `0 days 07:45:00`

- sleep_rating (float64): Numerical rating of sleep quality. Example: `0.65`

Notes

- The `Value` field stores JSON-like strings which should be parsed for specific values such as heart rate (`bpm`).

- Missing data in `screentime`, `expected_sleep`, and `sleep_rating` should be handled carefully during analysis.

- Timestamps are in Unix format and may need conversion to readable datetime.

Provenance

The Group Health (Sleep and Screen Time) Dataset was collected by the students at the Eindhoven University of Technology as part of a health monitoring study. Participants wore wearable health devices (Mi Band Smartwatches) that tracked biometric data, including heart rate, screen time, and self-reported sleep information. The dataset was compiled from multiple sessions of device usage over time the course of two weeks, with the data anonymized for privacy and research purposes. The original data was already in a standardized csv format and was altered for preprocessing purposes and analysis. This dataset is openly shared under a CC-BY 4.0 license, enabling users to reuse and modify the data while properly attributing the original creators