Graphs are one of the most powerful yet underutilized abstractions in Scientific Python. From social interactions and biological systems to text, infrastructure, and machine-learning pipelines, many complex phenomena can be understood not as tables, but as networks of relationships.

This talk introduces the core ideas of graph theory through a practical, Python-first lens, showing how real-world problems can be modeled as nodes and edges using NetworkX. We will build intuition around fundamental concepts such as paths, connectivity, components, and graph types, and then explore why structure matters: how different network topologies reveal hidden patterns that traditional data representations often miss. A central focus will be on centrality measures and what they actually mean in scientific contexts. Rather than treating them as abstract formulas, we will interpret centrality as influence, importance, flow control, or vulnerability, depending on the system being modeled. Through concrete examples, attendees will learn how to ask better questions of their data by reframing it as a network.

By the end of the session, participants will understand when and why graph-based thinking is useful, how NetworkX fits naturally into the Scientific Python ecosystem, and how graph theory provides a unifying language for analyzing structure, interaction, and complexity across disciplines.