Alertify – ML Based Disaster Detection System

Alertify is a machine learning–based disaster detection system designed to identify hazardous situations and trigger automated alerts.

Project Background

The base object detection model was initially cloned for experimentation. Due to suboptimal performance, the model was retrained and evaluated on a balanced dataset to improve detection accuracy and reliability.

My Contributions

• Trained the ML model for better performance
• Balanced the dataset for better class distribution
• Evaluated model using confusion matrix and performance metrics
• Integrated ML output with a real-time dashboard
• Implemented dynamic status indicators (Safe / Alert / Danger)
• Integrated automated email alert system

Tech Stack

• Python
• YOLO (Object Detection)
• OpenCV
• Flask
• HTML, CSS, JavaScript
• Google Colab (Model Training & Testing)

Repository Notes

• Datasets and trained model weights are excluded due to size constraints
• Training and inference logic is documented in the model notebook