MONITOR PhysiCube Dataset: Dataset of Physiological Parameters for Monitoring Stress in Garments Workers using PhysiCube Sensors

MONITOR: A Data-driven Intelligent MONITORring System to Improve Quality of Working Life

The project, MONITOR, aims to establish mobility of bilateral collaboration between Islamic University of Technology (IUT), Bangladesh, and Mälardalen University (MDH), Sweden in specific research and education domains of Science and Technology at an institutional level. MDH has been leading several projects to develop tools to classify physiological data for individualized classification of physiological sensor readings to monitor stress, fatigue, and mental load at work and home environments. The project was funded by Stiftelsen för internationalisering av högre utbildning och forskning (STINT) under the Grant Agreement No. 202100-2916.

This project has unfolded many benefits for both the partnering universities. It gave access to a wide variety of data regarding mental health-related parameters to the researchers. Given the data collection was done in Bangladesh, it was challenging for the researchers to extract correct and accurate data under harsh/unsafe conditions. This trained the researchers to collect data under real complex conditions. In this project, MDH provided resources and knowledge to update the already developed AI algorithms to support Garment workers’ mental health monitoring. IUT provided expert human resources to develop a sensor-setting framework and assisted in collecting and analyzing the data.

Northern Tosrifa Group (NTG): A LEED-certified garment factory in Bangladesh, provided the test environment and subjects to collect and validate the system. The group has 5 companies under its operation and all the companies manufacture and export their products to markets in North America and Europe. For the project MONITOR, Fashion Asia Limited (FAL) (one of the 5 companies under NTG) provided the necessary manpower and data collection environment for the project. 

About the Dataset

As part of the MONITOR project, an in-house sensor device, named PhysiCube, was developed for measuring the following physiological parameters-

Object Temperature(F): The skin temperature of the subject (measured in Fahrenheit) was recorded using a low-cost off-the-shelf temperature sensor.

Heart Rate Ear(BPM): The heart rate of the subject (measured in Beats per Minute) was recorded using a heart rate sensor wearable on the ear lobe.

Galvanic Skin Response (GSR): The Galvanic Skin Response (GSR), which measures the skin’s electrical conductivity in response to sweat gland activity in the fingers and palms was measured using a low-cost off-the-shelf GSR Sensor.

The dataset contains the timestamped values (as elapsed time, in the format '%MM:%SS') of all the aforementioned parameters for 24 subjects (29.17% Male, 70.83% Female, Age Group: 18-40 years) with written consent.

Data were collected in the following three phases-

Baseline Values: Values of the aforementioned physiological markers that reflect the corresponding values under normal/relaxed conditions. In other words, the stress-free values of these physiological markers. Example filename: C01_PF_base.csv.

After-Lunch Values: Values that were recorded after the lunch break. Example filename: C10_PM_after_lunch.csv.

Stress-Induced Values: Values that were recorded while inducing stress using various stress-inducing questionnaires available in the literature. Example filename: C10_PM_stress_induced.csv.

Files were organized in a way such that each subject had a directory. In each directory, files were named as '__.csv'. For example:

Base data of a Female subject C01 recorded using a PhysiCube sensor will have the name: C01_PF_base.csv

After Lunch data of a Male subject C10 recorded using a PhysiCube sensor will have the name: C10_PM_after_lunch.csv

N.B. : As part of the MONITOR project, another dataset titled, "MONITOR Shimmer Dataset: Dataset of Physiological Parameters for Monitoring Stress in Garments Workers using Shimmer Sensors", was simultaneously developed using Shimmer sensors, to measure the physiological data of the same subjects.