Skip to content

docs(examples): worked graph-exploration walkthrough (Cypher -> GDS -> embeddings)#73

Merged
colombod merged 1 commit into
mainfrom
docs/graph-exploration-example
Jul 14, 2026
Merged

docs(examples): worked graph-exploration walkthrough (Cypher -> GDS -> embeddings)#73
colombod merged 1 commit into
mainfrom
docs/graph-exploration-example

Conversation

@colombod

Copy link
Copy Markdown
Collaborator

Summary

A worked, end-to-end example of using context-intelligence as an analytics substrate over Amplifier session history. Demonstrates the full escalation path: from a plain-language question through Cypher graph queries → Neo4j GDS (project.cypher, knn, louvain, pageRank) → OpenAI embeddings for semantic intent clustering.

What's Included

  • graph-exploration-walkthrough.md — narrative walkthrough showing the method, discoveries, and the dual clustering insight (machinery vs. mission)
  • Five rendered diagrams — method escalation pipeline, agent co-occurrence communities (Louvain + PageRank), delegation star, GDS session-clustering transformation, and the key discovery (machinery-vs-mission orthogonality)

Corpus Scale

  • Team-shared environment: ~2,065 sessions / 42.4K tool calls
  • Private home-server environment: ~3,375 sessions / 56,653 tool calls / 44 agents / 2,598 delegations
  • Total prompts embedded: 3,199 first-prompts (OpenAI text-embedding-3-small, 1536-dim → k-means)

Note for Reviewers

This example contains real mined session content (workspace names and short prompt excerpts from a single-user corpus). Please review and flag if any anonymization is needed before merge.

Related Work

Pairs with the skill (newly landed on main — the reusable procedure) and the existing (the how-to reference).


Docs only. No code changes; conform validation unaffected.

…DS -> embeddings)

A worked, end-to-end example of using context-intelligence as an analytics substrate over
Amplifier session history: starting from a plain-language question and escalating through
Cypher graph queries -> Neo4j GDS (project.cypher, knn, louvain, pageRank) -> OpenAI
embeddings for semantic intent clustering. Includes five rendered diagrams (method
escalation, agent co-occurrence communities, delegation star, GDS session-clustering
transformation, and the machinery-vs-mission discovery). Pairs with the
context-intelligence-gds skill (the reusable procedure) and the existing
context-intelligence-exploration-guide (the how-to). Corpus: team-shared ~2,065 sessions +
private ~3,375 sessions / 56,653 tool calls; 3,199 first-prompts embedded.

🤖 Generated with [Amplifier](https://github.com/microsoft/amplifier)

Co-Authored-By: Amplifier <240397093+microsoft-amplifier@users.noreply.github.com>
@colombod colombod merged commit 5da391e into main Jul 14, 2026
8 checks passed
@colombod colombod deleted the docs/graph-exploration-example branch July 14, 2026 08:44
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant