City Hub Detection Powered by Smart Cyber-Physical Social System

The concept of smart cities has gained significant attention due to its potential to enhance citizens' lives by offering services like smart transportation, homes, and telecommunications. Hotspot analysis, a key aspect of spatial analysis, is particularly important for telecommunication companies seeking to identify areas with high communication activity.Cyber-physical-social systems (CPSS) have proven effective in detecting hotspots in smart cities. However, challenges such as big data storage, processing, accuracy, and robustness remain. To address these, we propose a smart CPSS model for hotspot analysis using telecom data.Our model consists of three layers, each with distinct functions. The data collection layer gathers raw Call Detail Record (CDR) data, which is then processed in the second layer. This stage includes data pre-processing, storage, and analysis, followed by the construction of a graph and social network analysis (SNA). Unlike conventional approaches, we utilize Eigenvector and k-shell as social network similarity measures, alongside Jaccard and cosine for social behavior assessment.Through SNA, we identify and rank the top ten hotspots based on key metrics. Using five-day data, we analyze variations in hotspot patterns and validate our results by comparing them with the original dataset. Autocorrelation and cross-correlation functions confirm the model's accuracy and robustness.