Readme for Unlabeled Empatica E4 Wristband Data (UE4W) Dataset

DOI 10.5281/zenodo.6898244

Lee B. Hinkle and Vangelis Metsis
Intelligent Multimodal Computing and Sensing (IMICS) Lab
Texas State University

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

The UE4W dataset was recorded over a period of approximately two weeks by a singe adult male.   We found that keeping accurate records and labeling were impractical so it was archived.  However, recent advances in representation and unsupervised learning have led us to look at it again.   The dataset contains over 250 hours of recordings using an Empatica E4 wristband.  In general the longer files are daytime recordings, the shorter ones are nighttime.   The tags indicate the beginning and end of an event change (e.g. getting up to go for walk) however the lack of a tag is not a meaningful indicator.

Each zip file contains multiple datafilee for each of the sensors on the E4 wristband.   The file name is based on the time and serial number of the wristband.   Several files are listed as wristband 00000 but the data seems OK.   There are a few anomalies due to operator errors and battery life/charging issues.

Contents:
15***_A01F11.zip - each of these files represent a recorded session, the format is per the download from the Empatica Connect site.

connect_plot_752***.jpg - screenshots of the Empatica graphs are provided for the first two sessions.

connect_table_session_list - a screenshot showing all of the files, a .csv version is provided as well.

load_ue4w_demo.ipynb - a Jupyter notebook that loads each of the files, generates a Pandas data frame, and shows a windowed plot of the total acceleration.