Analysing Social Media Images and Text for Disaster Response

Problem Statement

Disaster response teams require timely and accurate information to prioritize resources and save lives. Social media platforms like Twitter and Instagram provide valuable real-time data during disasters. However, manually processing and filtering large volumes of multimedia posts to extract critical disaster-related content is inefficient. This project aims to develop a system that automatically processes both images and text from social media posts to detect disaster-related information, enabling faster response times and resource allocation.

Problem Deliverables

The project delivers a comprehensive system capable of:

1. Detecting disaster-related objects and events in images.
2. Analyzing text posts to identify disaster-related information, such as calls for help or damage reports.
3. Integrating image and text classification models into a web application for real-time social media post processing.
4. Ensuring the system can handle multiple requests simultaneously, providing seamless user interaction and scalability.

Documentation

Link - Project Documentation

Installation

1. Clone the Repository

   git clone <repository_url>
   cd Analysing-Social-Media-Images-and-Text-for-Disaster-Response

2. Download Model Weights

   • Download the ResNet model weights from Google Drive link.
   • Place the downloaded file in the models folder.

3. Set Up Virtual Environment

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

4. Install Dependencies

   pip install -r requirements.txt

5. Run the Flask Server

   flask run

6. Access the Web Application

   • Open a web browser and go to http://127.0.0.1:5000 to interact with the application.
   • Project HomePage:

Usage

• Upload social media images and text posts via the web interface.
• The system will analyze the content and provide disaster-related information.
• Monitor the results and adjust the system settings as needed.

Contributing

• Fork the repository and make your changes.
• Submit a pull request with a detailed description of your modifications.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contact

For questions or feedback, please contact Rakshit Garg.