Profile segmentation: Clustering approach based on behavioral patterns extracted from mobile phone data

Abstract

The ubiquitous use of mobile devices in recent years has enabled learning more about human behavior. This study uses mobile phone data, point-of-interest data, and publicly available socioeconomic indicators to present an integrated analysis of individual socioeconomic and behavioral profiles in Lisbon. To effectively segment users and identify mobility patterns, we employ multiple clustering techniques tailored to different data types: varied density-based spatial clustering of applications with noise (VDBSCAN) detects meaningful places by identifying clusters of varying densities in mobile phone data, while K-Means and K-Modes cluster numerical and categorical data, respectively, to determine socioeconomic status (SES) and categorize work environments. Our analysis identifies four primary groups of individuals, each distinguished by different SES and behavioral tendencies, ranging from practical and family-oriented to lifestyle-driven patterns. The results show that SES strongly influences mobility behaviors and interactions with urban spaces, revealing disparities in housing access, services, and work environments between different population segments. These findings highlight the interaction between mobility, social habits, and economic conditions, offering key insights for urban planning and strategies to mitigate the impacts of gentrification. The proposed methodology can be applied to other cities to support inclusive policymaking, transportation planning, and spatial equity strategies, fostering more accessible and equitable urban environments.