Disaster-Management-SMS-Pipeline

Objective

The objective of this project is to aid the disaster relief process by making communication and the alerting process more efficient.

Dataset

The dataset used in this project for training and testing is taken from Figure Eight consisting of disaster messages along with their corresponding appropriate responses.

Model

The model used in this project is made using the concept of Natural Language Processing for classification with the help of PyTorch. Its job is to classify the disaster based on the message received and also visualise the worked data.

Website

This project uses a website as an interface in which users have registered accounts. The victim sends a message as a query(input) in the website which then passes it to the deployed model(using Flask) which classifies the type of disaster and then sends a notification to the registered account of the appropriate disaster relief agency. The website makes use of HTML, CSS and JavaScript for frontend and MongoDB, Flask and NodeJS for backend.

Drive Link for the Random Forest Model

https://drive.google.com/drive/folders/1wNbSg3F_4P77t6EhB5kfQgwMAVq8EY_V?usp=sharing