Data Collection Protocol

During a single data collection experiment, participants were exposed to eight different experimental session conditions that focused on various work environments, temperature levels, and humidity labels based on the ASHRAE scale. Each session was carefully designed with specific activities and varying humidity and temperature levels. For example, gymnastics activities were performed in both hot conditions (32°C temperature and 80% humidity) and warm conditions (25°C temperature and 60% humidity). Similarly, reading and typewriting activities were conducted under different temperature and humidity states. The duration of gymnastics activities was 10 minutes, while the other activities lasted for 15 minutes. To ensure a smooth transition between activities, participants were given a 5-minute break after completing the gymnastics activities, which were conducted under hot temperatures (32°C) and high humidity (80%) as well as under warm temperatures (25°C) and moderate humidity (60%). This break was implemented to prevent the influence of the intense physical exertion in hot and humid conditions on the subsequent experiments. Table 1.1, presents each experiment session condition elaborately.

Data Overview

The data for this challenge has been collected in a controlled environment to assess both the impact of the thermal environment on biological information and the consequences of work overlapping with mental or physical burdens. The goal is to understand how such overlap can contribute to more severe conditions, potentially leading to heatstroke.

Overall, the challenge dataset consists of data from a total of 36 participants over a span of  8 days. The training dataset contains 27 participants’ 6 days of data in varying temperatures and humidity levels for different work conditions, i.e., reading, typewriting, and gymnastics activities data focusing on hot thermal conditions, and we have set these activities on real-world circumstances. For example, elderly people reading/watching television at home will be associated with reading activities, office work classroom study will be aligned with reading and typing activities, and factory/outdoor labor requiring higher effort will be aligned with heavy work activities like gymnastics. The goal is to gain a better understanding of how individuals respond to high temperatures while performing different activities. Also, during the experiment, participants reported their subjective thermal sensation states every five minutes using a thermal assessment logging application that we designed and placed on a smart tablet.


During the experiment, participants were requested to wear Empatica E4 wristbands, which is a medical-grade wearable device that offers real-time physiological data that contain numerous sensors, including an Electrodermal activity (EDA) monitor sensor, a photoplethysmography (PPG) sensor (which measures the Blood Volume Pulse (BVP), which is a metric that may be used to determine heart rate variability to assess sympathetic nervous system activity and heart rate simultaneously at the same time), and a three-axis accelerometer (ACC), and an optical thermometer. Through Emapatica, we also collected participants' Inter-Beat Interval (IBI) signal, which refers to the amount of time that passes between each individual beat of the heart. Participants need to use extracted Heart Rate Variability (HRV) featured data along with a subjective evaluation to forecast individual thermal comfort labels.

Data Structure

For the challenge, the dataset has been divided into two subsets. The first subset contains data from 23 subjects across 6 days and contains all personal thermal comfort sensation labels. The second subset contains data from different 6 subjects across 4 days, and is not labeled. Participants must submit their results of forecasting labels in these 4 days, given past data on the second dataset using their models. 


Featured data contains following information:

Data Use
All participants may use the data free of charge. Download the dataset


Result Submission
Participants should forecast personal thermal comfort labels for different days. Please submit the forecast labels with subject ID and timeframe (also, required to submit their assessment code for an evaluation).

The submission file contains the columns detailed below.

Prizes
- The winning team will be awarded 100,000 jpy!
- The registration fee for the 1st and 2nd runner-up teams will be waived.
- Each of the participating teams will be awarded with participation certificate.