Sleep study comparison using the Apple Watch and the Philips Actiwatch
Fourteen healthy adults (9 males, 5 females) were recruited by word of mouth and direct approach to participate in a sleep study. The Queensland University of Technology Human Research Ethics Committee (#1800000242) approved all procedures, and all participants gave their signed consent prior participating the study.
This dataset contain the records of participants' movement during their sleep at home.
Participants were asked to wear Apple Watches and Actiwatches (Philips) on their non-dominant wrist for two consecutive nights. All wearable devices were then returned for data extraction. One participant forgot to charge the Apple Watch so we lost one night of data and were left with 27 nights from 14 participants.
This dataset was created as part of the PhD of Sirinthip Roomkham since 2012 through the school of Electrical Engineering and Computer Science, Queensland University of Technology.
Download and unzip the sleep dataset. The dataset contain 27 files.
The filename is based on participants and the nights of recording, as following examples:
- p02_n01_ready.csv : this is the Participant 2 (p02), and it is the first night of recording (n01)
- p02_n02_ready.csv : this is the Participant 2 (p02), and it is the second night of recording (n02)
Each file contain five columns:
- timestamp : date and time
- Actiwatch activity counts : the activity counts of Actiwatch Spectrum
- Actiware classification : the sleep classification 0= sleep, 1 = wake
- Apple Watch ENMO: Elucidean Norm Minus One
Geographical area of data collection
Cite this collection
Access the data
Creative Commons Attribution-NonCommercial 4.0 (CC-BY-NC)