In archaeological studies, networks are constructed using both actual observations and hypotheses about interactions, exchanges, organization, etc. This chapter describes this modeling process and outlines the main differences between models primarily based on data and those constructed using complex sets of hypotheses, as well as their impact on the subsequent analyses. It reviews a selection of measures, metrics, and properties taking roots in graph theory and commonly used to describe networks at different levels of aggregation, with a particular focus on their applicability on different types of networks. It emphasizes the core role of paths in network analysis as well as the impact of modeling hypotheses on the expected structure of a network. Finally, the chapter discusses differences and complementarity between classical data analysis and network based analysis, in light of the nature of archaeological data and of the associated modeling processes.