🫀 CardioSense

Towards Intelligent and Personalized Cardiovascular Health Monitoring

CardioSense is an ML and IoT-driven cardiovascular health monitoring system developed to demonstrate the integration of physiological sensing, machine learning–based risk analysis, and real-time web visualization with AI assistance.

The system collects cardiac signals using wearable sensors, processes and analyzes the data through a machine learning model, and presents personalized health insights via an interactive dashboard and AI assistant.

CardioSense is designed as a research and academic prototype aimed at early health awareness, system experimentation, and intelligent decision support rather than clinical diagnosis.

---

📌 Key Features

• Real-time acquisition of ECG, Heart Rate, and SpO₂ signals
• Wireless data transmission using ESP32 microcontroller
• Machine Learning–based cardiovascular risk prediction
• Interactive web dashboard with live vitals and waveform visualization
• AI-assisted chatbot for explanations and health recommendations
• Modular and extensible system architecture

---

🏗️ System Overview

Hardware Layer (Data Acquisition)

• MAX30100 – PPG signal, Heart Rate, and SpO₂ measurement
• AD8232 – ECG signal acquisition
• ESP32 – Sensor interfacing and Wi-Fi communication

Software Layer (Analysis & Interaction)

• Flask Backend – API handling, data processing, ML inference
• Machine Learning Model – Logistic Regression–based risk prediction
• Web Dashboard – Real-time visualization and monitoring
• Chatbot Module – AI-based explanations and recommendations

---

🤖 Machine Learning Model

• Algorithm: Logistic Regression
• Input Features:
  • Heart Rate (HR)
  • SpO₂
  • Heart Rate Variability (HRV)
  • Systolic Blood Pressure
  • Diastolic Blood Pressure
  • Age
• Output: Probabilistic cardiovascular risk score categorized as Low, Medium, or High

The model is trained using a synthetic dataset that reflects realistic physiological ranges and inter-feature relationships, allowing controlled experimentation and evaluation.

---

📊 Dashboard Capabilities

• Live ECG and PPG waveform visualization
• Continuous display of HR, SpO₂, HRV, and Blood Pressure
• Color-coded cardiovascular risk indicator
• User History of Heart Rate and SpO₂
• AI-assisted health insights based on short-term trends
• Integrated AI chatbot for explanations, guidance, and user interaction
• Session-based user monitoring

---

📁 Project Structure

CardioSense/
│
├── app.py # Flask backend server
├── cardiosense_dashboard.html # Frontend dashboard
├── cardio_risk_dataset.csv # Training dataset
├── cardio_risk_model.pkl # Trained ML model
├── train_cardio_risk_model.py # Model training script
├── requirements.txt # Python dependencies
├── static/
│ └── css/ # Stylesheets
├── package.json # Frontend dependencies
├── tailwind.config.js # Tailwind CSS configuration
└── README.md # Project documentation

---

🚀 How to Run the Project

1️⃣ Clone the Repository

git clone https://github.com/ashwinraj8090/CardioSense.git
cd CardioSense

2️⃣ Install Python Dependencies

pip install -r requirements.txt

3️⃣ Start the Flask Backend

python app.py

4️⃣ Launch the Dashboard

Open cardiosense_dashboard.html in a browser (Live Server recommended)

Ensure the Flask server is running before starting monitoring

---

🔑 API Key Configuration

This project uses an external AI service for chatbot-based health explanations and recommendations. API keys are not included in this repository for security reasons.

📍 Where to add your API key

Add your API key in the frontend JavaScript section inside:

cardiosense_dashboard.html

Look for a placeholder similar to:

const AI_API_KEY = "YOUR_API_KEY_HERE";

Replace "YOUR_API_KEY_HERE" with your own valid API key.

---

📡 IoT Integration

The ESP32 continuously sends sensor readings to the Flask backend using HTTP-based REST APIs. The system supports:

• Simulated data mode for testing and development

• Live IoT mode for real-time sensor integration

---

🔮 Future Enhancements

• Clinical validation with real-world patient data

• Advanced ML/DL models for improved prediction accuracy

• Mobile application support

• Cloud-based deployment and scalability

• Secure authentication and encrypted data storage

---

📜 License

This project is licensed under the MIT License.