Intelligent Urban Activity Analysis Using Call Detail Records and CPSS Model

Rapid urbanization of the 21st century has shaped cities into intricate ecosystems, requiring intelligent systems for operational resource mobilization and citizen safety. Traditional methods of monitoring clusters and usage spaces of information communication technology resources in urban environments are unable to keep pace with the real-time dynamic, dependent nature of those urban environments by relying on static data analysis techniques. This paper introduces a Smart Cyber-Physical-Social System this able to detect urban activity clusters using Call Detail Records, and because of the sheer volume of recently escalated telecommunication networks, we were able to embrace a Smart Cyber-Physical-Social System (CPSS). The CPSS framework as proposed in the paper, is a three-layer system: Data Collection; Data Preprocessing; Social Network Analysis. The CPSS framework employs a hybrid Eigenvector and k-shell centrality, combined with Jaccard and cosine similarity as needed to provide an accessible means to ensure accuracy and reliability in measuring clusters. The CPSS provides a JSP web interface to visualize real time information, and the Smart CPSS separates cluster information into Safe, Moderate, and Danger. In a 5-day dataset experimental validation, the model and JSP interface achieved 95% accuracy with respect to classification, and reduced false positives, and improved robustness definitions of dangerous clusters from 100m to any reasonable distance. As a smart cyber-physical-social system, we offer an intelligent tool for urban planning, optimizing communication use, engaging public safety issues and enabling smart city management with a commitment towards improving resilient and connected urban ecosystems.