Modern fast-paced lifestyles have significantly increased stress-related health risks. Traditional assessment methods—such as clinical surveys or periodic checkups—are often subjective, episodic, and fail to capture sudden physiological spikes. This project bridges the gap by delivering an end-to-end, automated, and non-invasive Real-Time Stress Detection and Analysis System.

​The system collects continuous sensor streams (Heart Rate, \text{SpO}_2, Ambient Temperature, Humidity, and Fall Detection) and enhances this raw data using biomedical calculations. It computes critical health metrics, including:

• ​Blood Pressure (Systolic & Diastolic)
• ​Pulse Pressure (PP) & Mean Arterial Pressure (MAP)
• ​Heat Index & Oxygen Delivery

​This multi-parameter strategy ensures that cardiovascular, thermoregulatory, and respiratory stress responses are captured simultaneously.

​2. Hybrid Ensemble Intelligence

​Instead of relying on a single classifier, the backend employs a Soft-Voting Ensemble Model that combines:

• ​Random Forest (handles non-linearity and noisy data)
• ​Support Vector Machine (SVM) (provides strong boundary separation)
• ​Neural Networks (captures deep feature interactions)

​This hybrid framework minimizes bias, improves prediction accuracy, and ensures reliable classification across varied patient profiles.

​3. Full-Stack Monitoring Ecosystem

​Built on a lightweight Flask web framework, the system includes:

• ​Interactive Live Dashboard: Built with Chart.js to show real-time line graphs for heart rate and \text{SpO}_2, along with dynamic color-coded indicators (Normal, Warning, Critical).
• ​Automated PDF Reports: Uses ReportLab to generate formal medical reports after every prediction cycle.
• ​Database & Admin Control: Saves clinical data to a SQLite database, enabling time-series analysis and administrative patient management for healthcare providers.