1  Introduction

Descriptive network analysis is the task of summarizing the structure of an observed network directly from the data, without assuming any probabilistic model for how the ties came to be. The quantities you compute here are properties of the network you have in front of you: how many nodes and edges it contains, how far apart they are, who sits at its center, which regions hang together. The visualization methods covered later in the book help you communicate these structures, and the inferential models help you test hypotheses about them, but both build on the vocabulary introduced in this part.

These descriptors are the natural first step in any network study. Before drawing a layout or fitting a model, it pays to know a few basic numbers: the size and density of the network, whether it is connected, how clustered it is, which nodes are most prominent. The answers shape every later choice. A sparse, fragmented network calls for different visual encodings than a dense one, and a directed network with strong reciprocity calls for different models than an undirected one. Descriptive statistics are also what you use to validate a fitted model: a good model should reproduce the descriptors you care about.

The chapters in this part move from global to local, then to specialized structures, and finally to a more abstract view. Basic Network Statistics introduces the whole-network descriptors: size, density, distance, components, transitivity, reciprocity, and the dyad and triad census. Centrality then shifts from the network as a whole to the prominence of individual nodes. Cohesive Subgroups takes the intermediate view, covering cliques, community detection, blockmodeling, and core-periphery structure. From there, three chapters cover network types that need their own toolkit: Two-Mode Networks, Signed Networks, and Ego Networks. The part closes with Entropy Analysis, which offers an information-theoretic lens on the dependence structure among variables attached to a network.

Taken together, these descriptors form the working vocabulary used throughout the rest of the book. Many later chapters assume you are comfortable reading a density, a centrality score, or a community assignment without a second thought, so the ground covered here is foundational rather than preliminary.