Exploring human mutuality in cyber and physical spaces using mobile big data and network analysis

Abstract

Partially due to the limited access to datasets of human activities in cyber and physical (online and offline) spaces, the exploration of weak human interactions, defined as human mutuality in this work (i.e. co-location in physical space, and co-domain in cyber space) and their networks in the two spaces have been constrained to some extent in recent years. To bridge this gap, this study establishes a unified framework for directly comparing individual-level human mutuality networks across physical and cyber spaces, based on large-scale Uniform Resource Locators (URLs) data from tens of thousands of users in Jilin, China. Within this framework, human mutuality networks are constructed with users as nodes and mutuality events as edges, based on shared locations or shared website visits. The networks are systematically analyzed through three dimensions: fundamental network properties (such as clustering coefficient and average shortest path length), degree and strength distributions, and community structures. The results show distinct structural differences between the two spaces. Cyber space displays a significantly shorter average shortest path length (2.4) than the physical space (7.6), suggesting faster information transmission and the potential to alleviate digital inequalities by accelerating access to resources. Both networks present heavy-tailed degree distributions, indicating heterogeneous structures shaped by a few highly connected individuals. Furthermore, while physical space exhibits numerous small communities with strong local clustering, cyber space contains fewer but larger communities, with weaker local cohesion. This reduced local clustering may increase the risk of rapid misinformation diffusion. Additionally, the formation of cyber communities based on shared online behaviors reveals potential socioeconomic similarities among users despite differences in their physical attributes. Together, these insights offer a foundation for understanding human interactions across hybrid spaces and inform strategies for managing cyber and physical social dynamics.