11_social_network

Example 11 — Social network analytics

What it demonstrates

A small end-to-end social-network workload over a labelled property graph: attach labels (User, Verified) and typed properties (age) to people, freeze the graph into an immutable CSR snapshot, then run three analytics over it — PageRank influence ranking, Leiden community detection, and a manual two-hop friend-of-friend recommendation walk.

Domain / scenario

Seven people form an undirected friendship graph. Each person is a User node carrying an age property; the three accounts that are verified also carry a Verified label. Friendships are unweighted, symmetric edges:

alice — bob      bob   — dave
alice — carol    carol — erin
alice — erin     carol — frank
dave  — grace    erin  — grace

The analytics then read this graph three ways: who is most central (PageRank), which clusters the friendships form (Leiden), and who alice should befriend next (the people two hops from her who are not already her friends).

How to run

go run ./examples/11_social_network

Expected output

Influence (PageRank):
  carol    0.1855
  alice    0.1786
  erin     0.1786
  dave     0.1294
  bob      0.1270
  grace    0.1270
  frank    0.0740

Communities (Leiden):
  community 0: [alice bob dave]
  community 1: [erin grace]
  community 2: [carol frank]

Friend-of-friend recommendations for alice:
  -> dave
  -> frank
  -> grace

The output is byte-stable: PageRank ties (alice/erin and bob/grace share a score) are broken by name, the Leiden clusters are printed in sorted cluster-id order, and the recommendations are sorted by shared-friend count then name.

Key APIs

• graph/lpg.New / Graph.SetNodeLabel / Graph.SetNodeProperty — build the labelled property graph and attach typed age values via lpg.Int64Value.
• graph/lpg.Graph.AddEdge — add the undirected friendship edges.
• graph/csr.BuildFromAdjList — freeze the live adjacency list into an immutable CSR snapshot for analytics.
• search/centrality.PageRank / centrality.DefaultPageRankOptions — rank users by influence.
• search/community.Leiden / community.DefaultLeidenOptions — detect communities; the result's Community slice maps each NodeID to a cluster id.
• graph/adjlist.AdjList.Mapper (Resolve) / AdjList.Neighbours — translate NodeIDs back to names and walk the live adjacency list for the friend-of-friend recommendation.

Further reading

• search/centrality — PageRank and other centrality measures
• search/community — Leiden and community-detection algorithms
• graph/lpg — the labelled property-graph builder used here
• Example 08 — PageRank — PageRank in isolation
• Example 09 — Leiden — Leiden community detection in isolation
• docs/examples-standard.md — the standard every example follows