A graph-based engine to assess organizational power structures for hidden risks, miscommunication, and potential corruption.
Upload a DOT graph, get a transparency index and structural diagnostics.
Power Is Structure is an open-source engine that represents organizations as graphs of authority and communication.
It analyzes them to detect corruption-prone patterns, information flow asymmetry, and control bottlenecks.
The core idea is simple:
Corruption is not random — it emerges from structure.
By modeling hierarchies as graphs and identifying specific anti-patterns, we calculate a Power Transparency Index,
a quantitative estimate of how structurally robust (or fragile) an organization is.
- 🏛️ Government ministries and civil institutions
- 🏢 Corporations and enterprise departments
- 🧩 NGOs and distributed organizations
- 🧠 Research in governance, transparency, and ethics
- 🧪 Experimental simulations and academic analysis
Example Outcomes:
- 🔍 Discover hidden authority loops before audits
- 📉 Quantify structural risks to guide reorganization
- 🛡 Support transparency efforts with measurable metrics
Submit a graph in DOT format describing authority and reporting links.
The engine parses your graph and detects structural issues:
⚠️ Opacity Triangles – filtered one-way chains of control- 🚨 Control Monopolies – unsupervised central managers
- 🕸️ One-way Subgraphs – no feedback paths upward
- 📉 Cascade Weakness – disconnected top management
- 🔀 Responsibility Conflict – multiple managers with no clear hierarchy
You receive a structured JSON report:
- ✅ Transparency Index (0–100)
- 🧠 Information Summary – key metrics, influencer roles, isolated nodes
- 🛑 Detected Issues – matched patterns and risk assessments
digraph structure {
A [label="CEO"]
B [label="Deputy"]
C [label="Staff"]
C -> B
B -> A
}{
"transparency_index": 76.4,
"information": {
"total_nodes": 17,
"total_edges": 28,
"hierarchy_levels": 4,
"top_influencers": [
{ "id": "ID001", "name": "Deputy Director", "incoming_edges": 8 },
{ "id": "ID002", "name": "HR Head", "incoming_edges": 6 }
],
"isolated_nodes": [
{ "id": "ID017", "name": "Consultant" }
],
"potential_risk_nodes": [
{ "id": "ID005", "name": "Middle Manager", "reason": "Single upward path, no cross-validation" }
]
},
"issues": [
{
"type": "Opacity Pattern",
"description": "Node ID005 has exclusive access to superiors and subordinates",
"nodes_involved": ["ID004", "ID005", "ID006"]
},
{
"type": "Responsibility Imbalance",
"description": "Node ID001 has 8 subordinates, Node ID007 has none",
"nodes_involved": ["ID001", "ID007"]
},
{
"type": "Disconnected Role",
"description": "Node ID017 is not connected to the graph",
"nodes_involved": ["ID017"]
}
]
}Organizational abuse doesn't stem from individuals — it stems from unchecked structures.
By making those structures visible and measurable, we bring accountability to design itself.
Think of it as linting for hierarchies.
- 📜 Manifesto of Structural Transparency
- 🧩 Structural Pattern Catalog
- 📄 API Documentation (soon)
- 🧪 Try Online (coming soon)
This project is released under the Mozilla Public License 2.0 (MPL-2.0).
You’re free to use, extend, and integrate the code.
Changes to the core logic must be published under the same license.
Created by Leonid Baranov,
software architect and organizational structure researcher.
Inspired by real-world patterns encountered in both post-Soviet and Western governance models.
- Core index calculation engine
- Anti-pattern detection
- Graph visualization patterns
- Public REST API
- Graph playground (web)
- HR/audit system integrations
- Plugin & pattern registry
Contributions are welcome!
- 📦 Submit new graph patterns
- 🧠 Improve detection heuristics
- 📊 Build visual tools
- ✍️ Share anonymized real-world examples
See CONTRIBUTING.md for details.