Digital Echoes: Empowering Crisis Management and Urban Resilience Strategies with Real-Time Social Media Insights

This paper proposes an innovative approach for decision-makers, introducing a web application that employs a microservices architecture to analyze data gathered from social networks. In the context of smart cities, citizens act as "sensors", providing real-time insights through their online activities. By sharing locations and posting content, they generate data that can be analyzed using supervised machine learning algorithms and natural language processing techniques to identify significant urban events. The application was trained on a pre-labeled dataset from X, utilizing various machine learning models and NLP preprocessing techniques to achieve high accuracy in message classification. The dataset comprises texts with keywords related to disruptive events, such as fires, floods, heatwaves, and more severe contemporary issues like terrorism, war, or epidemics, labeled as either disruptive or neutral. Comprehensive testing was conducted using X's API, focusing on acquiring messages from specific areas. This approach can enable more proactive city management and timely resource allocation, improving overall crisis response and urban planning.