# Network Analytics

> Network Analytics

- Skill: `dzhokhov/network-analytics-2` (Agent Skill, multi-file: 7 files)
- Install (CLI): `npx skillmds@latest add dzhokhov/network-analytics-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/dzhokhov/network-analytics-2/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- Author: dzhokhov (https://skillmd.com/u/dzhokhov)
- Updated: 2026-09-21
- Page: https://skillmd.com/skills/dzhokhov/network-analytics-2

---


# Network Analytics


> Navigation: [Skill index](../README.md)
## Overview

Transform chaotic mentions and interactions into a working map of the network of influence.

Skill is responsible for the `normalize entities -> update graph -> compute metrics -> outreach plan` phase.

## Workflow

### 1) Normalize Entities

Use the circuit from
[graph-schema](./references/graph-schema.md).

Substances:
- `Person`
- `Organization`
- `Event`
- `Topic`

### 2) Update Edges

Maintain typed connections:
- `co-speaker`
- `co-organizer`
- `mentioned_with`
- `works_at`
- `introduced_by`
- `attended`

For each rib, fix:
- `source`
- `date`
- `confidence`

### 3) Compute Metrics

If you have graph analytics tools, consider:
- `degree centrality`
- `betweenness centrality`
- `eigenvector centrality`
- `communities` (Louvain/equivalent)

If tools are not available, use heuristics and explicitly mark
`method=heuristic`.

### 4) Interpret

Use the rules of
[metrics-and-prioritization](./references/metrics-and-prioritization.md):
- bridges between clusters,
- the core of each domain,
- New growing clusters.

### 5) Build Outreach Plan

Give me:
- 5 key bridges
- 3 clusters/domains
- 10 priority outreach steps (to whom, through whom, why)

## Output

1. Graph update summary
2. Top bridges (betweenness)
3. Community map
4. Outreach backlog (10 items)

Each outreach item:
- `target_person`
- `via`
- `reason`
- `suggested_message_angle`
- `expected_outcome`

## Quality Gates

- No rib without a source or date.
- There are no central conclusions without specifying the method of calculation.
- Outreach recommendations are tied to a specific goal, not a general “get to know.”

## Anti-Patterns

- Do not mix names without proof of identity.
- Do not confuse frequent mentions with real influence.
- Do not draw conclusions about the cluster based on 1-2 connections.

## References

- [graph-schema](./references/graph-schema.md)
- [metrics-and-prioritization](./references/metrics-and-prioritization.md)