Investigating the relationship between physiological stress and environmental factors through Data Science, the Internet of Things and DIY wearables

Keywords: microclimate, Internet of Things, Data Science, physiological stress, wearables
Session Type: Virtual Paper Abstract
Day: Friday
Session Start / End Time: 2/25/2022 03:40 PM (Eastern Time (US & Canada)) - 2/25/2022 05:00 PM (Eastern Time (US & Canada))
Room: Virtual 5

Abstract

In this paper we will share early work investigating the intersection of local microclimate, affect, and human physiological response.

Currently, there is little research on the relationship between microclimate and humans’ emotions and health and – to the best of our knowledge – there are no readily accessible devices which can measure both microclimate and biometric data within a single unit.

We set out to build and programme a device that can measure microclimate and biometric data so that the data collected can be used to improve the environment and well-being of humans and the community.

The home-made device was built from readily available materials and consists of five sensors and a motherboard to measure: noise level, light intensity, amount of dust, carbon dioxide concentration, temperature, humidity and pressure.

A random forest regression model was trained on 80% of data set with the microclimate factors as input variables. The model was subsequently tested on remaining 20% of data set. Excel was used to store the accuracy and factors’ significance of the model.

Regression models for heart rate (HR), HR variability, short term HR variations, body temperature, sleep score and stress are accurate as R2 is higher than 0.3.

Most recently, we have acquired a pair of low-cost Arduino-compatible, 8-channel neural interface devices which can be used to sample brain activity (EEG). We hope to use data obtained from these devices as a third category of dataset to complement the existing physiological and local environmental datasets we already have access to.