Convolutional Neural Networks for Analysing Social Media Data to Improve Disaster Response and Crisis Management

Disasters, both natural and man-made, present formidable challenges to response and management activities. The development of social media has led to the creation of vast volumes of real-time data, which offer helpful insights for efficient crisis management. The data, however, is unstructured and noisy, and extracting meaningful information is challenging. This study investigates the application of Convolutional Neural Networks (CNNs) in social media data analysis to improve disaster response and crisis management. Utilizing the potential of deep learning, specifically CNNs, this study focuses on classifying posts related to disasters from various social media platforms. The paper elaborates on the data collection process, preprocessing methods, CNN model architecture, and evaluation metrics. Experimental result confirm that the CNN model is superior to traditional Machine Learning (ML) techniques in terms of efficiency and accuracy. The suggested model achieved an accuracy of 96.13%, precision of 92.08%, recall of 90.25%, and F1-score of 95.16%. The study identifies the potential of CNNs in real-time disaster response systems and provides suggestions for future enhancements in utilizing AI-driven solutions for crisis management.