A machine learning-powered web application that predicts and optimizes survival rates for out-of-hospital cardiac arrest patients. The system analyzes intervention, hospital, and AED location data to suggest optimal placements for new AEDs.

• WebsiteAED-Loacation-Optimization-in-Belgium
• StackPython, Dash, Plotly, XGBoost, Scikit-learn
• PlatformWeb
• FeaturesML Prediction, Interactive Maps, Real-time Optimization

Project Background

Seconds are critical in out-of-hospital Sudden Cardiac Arrest (SCA) cases. Early defibrillation is key in the 'Chain of Survival'. Public-access AEDs (Automated External Defibrillators) enable bystanders to provide rapid cardiac defibrillation, significantly improving survival chances.

Key Features

• • Interactive visualization of AED locations, hospitals, and patient data
• • Real-time survival rate predictions using machine learning
• • Interactive AED placement optimization
• • City-wise mortality rate analysis and trends

Technical Implementation

The project utilizes XGBoost for survival rate prediction, achieving an AUC score of 0.613. Features include distance to nearest AED/hospital, temporal data, and geographical information. The web interface, built with Dash and Plotly, enables real-time visualization and interaction.

Application Pages

• • Project Main Page: Overview and navigation
• • AED Optimization: Interactive map for optimizing AED placement
• • Monthly Mortality Analysis: City-wise mortality trends
• • Yearly Analysis: Long-term mortality rate patterns