Social Media Analytics Using AI-Enhanced Edge-Fog Computing

The exponential growth of social media data, particularly from platforms such as Twitter, has presented significant challenges for real-time processing, storage, and analysis, especially in environments where low-latency and context-sensitive analytics are critical. Traditional cloud-based systems struggle to meet these demands due to bandwidth limitations and high latency. This research introduces a hybrid edge-fog computing architecture enhanced by lightweight Artificial Intelligence (AI) models, including DistilBERT and LSTM, designed for real-time social media analytics. The study evaluates the performance of this hybrid system across three deployment configurations, edgeonly, fog-only, and hybrid edge-fog, using real-world Twitter datasets. The findings reveal that the hybrid edge-fog setup outperforms the other configurations, achieving an F1 score of 88.1%, an AUC-ROC of 0.93, and processing 45 tweets per second. This configuration demonstrated optimal trade-offs between latency, energy efficiency, and prediction accuracy. The research validates the potential of integrating AI with distributed edge-fog computing for scalable and agile social sensing applications, particularly in scenarios with limited connectivity or computational resources. These results provide valuable insights for the development of more efficient and adaptive systems for real-time social media analytics, with broader implications for IoT, smart cities, and other real-time data-driven applications.