From Variables to Relationships: The Role of Networks and Knowledge Graphs in Data Analytics

Abstract

Statistical methods have long provided the foundation for extracting knowledge from data. While traditional analyses primarily focus on variables and attributes associated with individual entities, many contemporary datasets also contain valuable information embedded in relationships, interactions, and dependencies among entities. The increasing availability of relational data has stimulated the development of network-based analytical frameworks that extend and enrich classical statistical perspectives. Community detection, link prediction, and multilayer network analysis provide powerful tools for exploring structural patterns and interaction dynamics in complex systems. Knowledge graphs further enhance relational representations by incorporating semantic information, enabling richer knowledge discovery and reasoning. Integrating attribute-based and relationship-centric perspectives offers a broader framework for understanding interconnected data and highlights the growing role of networks and knowledge graphs in the evolving landscape of data analytics.