How to Map a Social Network

Individual profiles tell you about people; the connections between them tell you about a network. Link analysis turns scattered SOCMINT into a graph that exposes clusters, brokers, and the accounts that actually hold a community together. This method builds that graph responsibly.

What you'll need

• Account and network tools from the Social Media category
• People and organization sources from the People & company category
• A note-taking or graphing approach to hold nodes (accounts) and edges (relationships)

Steps

1. Define the seed and the boundary. Start from one or a few known accounts and decide, before you begin, how far out you'll go (one hop? two?) and what counts as a connection — following, mutuals, replies, shared groups. Without a boundary the graph explodes.
2. Collect connections consistently. For each account, record the same relationship types every time so edges are comparable. Note direction (who follows whom) and, where you can, strength (frequency of interaction).
3. Pivot outward one layer at a time. Expand from the seeds to their connections, then stop at your boundary. Resist the urge to chase every interesting node — scope creep ruins the analysis.
4. Build the graph. Represent each account as a node and each relationship as an edge. Even a simple diagram reveals clusters (tight groups) and bridges (accounts linking otherwise separate clusters).
5. Read the structure, not just the size. The most-followed account isn't always the most important. Look for brokers who connect clusters, and for accounts whose removal would fragment the network — those are the real pressure points.
6. Corroborate before you conclude. A connection online is not proof of a real-world relationship. Weigh the evidence, note confidence, and separate "linked on a platform" from "associated in reality".

Common pitfalls

• No boundary. Expanding endlessly produces an unreadable hairball and buries the signal.
• Counting followers as influence. Reach and structural importance are different; brokers with modest followings often matter more.
• Reading a platform link as a real relationship. People follow strangers, bots inflate edges, and shared groups aren't friendships — say what the data actually supports.

Verify your result

You have mapped the network when you can point to its clusters, its bridging accounts, and its structurally central nodes — each backed by consistently collected connections and a stated confidence, not just a count of followers.