Repository for the Emognition Wearable Dataset 2020

Description

This is a supplementary repository for the Emognition Wearable Dataset 2020 and article titled Emognition dataset: emotion recognition with self-reports, facial expressions, and physiology using wearables.
The repository contains several jupyter notebooks with data manipulations and visualizations. The code uses only physiological and questionnaire data - none of the code requires video data.

Necessary libraries are in requirements.txt. The helpers.py script contains some general methods and constant values of the dataset.

Each notebook contains different analysis of the dataset:

• examination_signal.ipynb plots data of a selected participant during the study and introduce processing code for Samsung Watch BVP
• muse_quality.ipynb provides data on quality of recorded EEG signals (time on head and quality of signal - provided by the device). It allows
• questionnaires_analysis.ipynb provides a short analysis of control and emotion questionnaires
• skips_delays.ipynb contains analysis of skips (shorter duration) and delays (longer duration) of recorded sessions for washout and stimuli clips. All skips and delays were computed using signals recorded with the Empatica E4, as it was connected directly to the device used for elicitation of emotions.

Dataset access

The use of the Emognition dataset is limited to the academic research purposes only. The data will be made available after completing the End User License Agreement (EULA). The EULA is located in the dataset repository. It should be signed and emailed to Emognition Group at emotions<at>pwr.edu.pl. The mail has to be sent from an academic email address associated with the Harvard Dataverse platform account.

Project configuration

• Install Python (at least 3.7 version is recommended),
• Setup your environment (venv, conda, etc.) and install dependencies from requirements.txt (tested on miniconda3 for windows, 05.05.2021)
• Set path to the unzipped dataset in config.ini file. This path will be used in every notebook.

Reference

If you use the dataset or re-use this work, please cite:

@article{saganowski2022emognition,
  title={Emognition dataset: emotion recognition with self-reports, facial expressions, and physiology using wearables},
  author={Saganowski, Stanis{\l}aw and Komoszy{\'n}ska, Joanna and Behnke, Maciej and Perz, Bartosz and Kunc, Dominika and Klich, Bart{\l}omiej and Kaczmarek, {\L}ukasz D and Kazienko, Przemys{\l}aw},
  journal={Scientific data},
  volume={9},
  number={1},
  pages={158},
  year={2022},
  publisher={Nature Publishing Group UK London}
}