A CPSS-Driven Framework for Dynamic Urban Hotspot Detection Leveraging Socio-Spatial and Social Intelligence

With integrated services like automated housing, intelligent transportation, sophisticated communication networks, and effective infrastructure, smart cities are being created more and more to improve urban living. Finding locations with unusually high communication activity known as telecom hotspots is a major challenge in this field. For telecom companies looking to maximize network performance and resource allocation, identifying these areas is essential. Despite the potential of Cyber-Physical-Social Systems (CPSS) for this task, problems with large-scale data management, computational efficiency, and result reliability still exist. This work presents a sophisticated CPSS framework for hotspot detection using telecom datasets. Three functional layers make up the architecture: a data processing layer that handles data cleaning, storage, and analytical modeling; a hotspot analysis layer that conducts graph-based social network analysis; and a data acquisition layer that gathers raw Call Detail Records (CDRs). The model combines behavior-based metrics like cosine similarity and Jaccard similarity with network-based metrics like Eigenvector centrality and k-shell decomposition to improve detection accuracy. A five-day telecom dataset is used to assess the system and look at how communication patterns change over time. The robustness and accuracy of the system are confirmed by validating the identified hotspots using statistical techniques like autocorrelation and cross-correlation. Experimental findings demonstrate that the proposed approach correctly pinpoints regions of high activity, offering a scalable means of examining communication in smart city environments.