A Systematic Review of Social Media and Sensor Data Potential for Real-Time Regional Resilience Assessment Using Machine Learning

This study presents a systematic review of research utilizing social media and sensor data, combined with machine learning models, for real-time regional resilience assessment. Following the PRISMA ۲۰۲۰ guideline, a search in Scopus, Web of Science, and IEEE Xplore (۲۰۱۸–۲۰۲۵) yielded ۳۷ relevant articles. Findings reveal three main categories: (۱) social media data primarily for sentiment analysis, altruistic response identification, and regional disparity assessment; (۲) sensor and IoT data for infrastructure monitoring and early warning; and (۳) machine learning models such as Random Forest, LSTM, and deep learning for integrating these data sources. Key gaps include the lack of a standardized theoretical framework for "real-time resilience," challenges in data fusion, privacy concerns, and model interpretability. A conceptual model is proposed to guide future research toward a unified socio-technical resilience assessment.

کلیدواژه ها:

Regional resilience ، community resilience ، social media data ، sensor data ، machine learning ، real-time assessment ، systematic review