Data analysis and visualization

Abstract

Obesity and overweight have become a global issue and one of the most pressing concerns. According to the World Health Organization, nearly 2 billion persons are overweight, with 650 million obese [1]. As a result, there has been a rise in the number of people discussing weight loss. However, unintentional weight loss can occur, where a person loses more than 5% of their body weight in 6 to 12 months without actively attempting to [2]. Many who experience unintentional weight loss do not realize it. Such unintentional weight loss could be a symptom of a potentially fatal major health condition, thus, it is critical to detect it. To address this issue, this project aims to introduce unobtrusive health monitoring which uses ambient sensor technology to collect human health-related data without disrupting their daily life [4]. This project has developed a doormat prototype where human weight data is collected by simply stepping on it without having to stop. Models that were utilized include Random Forest Regression and CatBoost Regression to train the data collected to build a prediction model. This model will be integrated into a web application where it can predict human weight and send a warning notification whenever it detects an occurrence of unintentional weight loss.