Exploring online public survey lifestyle datasets with statistical analysis, machine learning and semantic ontology

Area of Science:

• Public Health
• Mental Health Research
• Data Science

Background:

• Lifestyle diseases are a major global health concern, with lifestyle factors heavily influencing depression.
• The COVID-19 pandemic exacerbated determinants of depression, increasing the need for research into contributing factors.

Purpose of the Study:

• To identify lifestyle and demographic factors associated with depression symptoms in India during the COVID-19 pandemic.
• To develop and showcase a methodology for collecting and analyzing online survey data for mental health research.

Main Methods:

• An anonymous online survey collected data from 1,767 participants in Kolkata, India.
• Statistical methods including principal component analysis (PCA), analysis of variance (ANOVA), and machine learning algorithms like support vector machine (SVM) and k-means clustering were used.
• Local Interpretable Model-agnostic Explanations (LIME) and OWL ontology were employed for model interpretation and data representation.

Main Results:

• A support vector machine (SVM) model achieved 96% accuracy in classifying depression-related factors.
• The study identified significant correlations between lifestyle, demographic variables, and depressive symptoms.
• The online survey methodology proved effective for data collection, visualization, and cost-efficiency.

Conclusions:

• Lifestyle and demographic factors are significant contributors to depression, particularly during health crises like the COVID-19 pandemic.
• Scalable online surveys offer a valuable tool for understanding and mitigating depression risk factors.
• The study provides a robust pipeline for future mental health research using digital data collection methods.