Organizational Network Analysis (ONA) Platform

A comprehensive Python application for analyzing organizational networks using machine learning and graph analytics. This platform transforms digital exhaust (emails, Slack messages, calendar data, code commits, etc.) into actionable insights about organizational structure, collaboration patterns, and network dynamics.

Features

• Multi-source Data Ingestion: Supports email, Slack/Teams, calendar, documents, code repositories, and HRIS systems
• Graph Construction: Builds weighted, temporal organizational graphs with sophisticated edge weighting
• Network Metrics: Computes centrality measures (degree, betweenness, eigenvector, closeness), structural holes, and core-periphery analysis
• Community Detection: Multiple algorithms (Louvain, Infomap, Label Propagation, SBM)
• Machine Learning: Graph Neural Networks (GCN, GAT), Node2Vec embeddings, link prediction
• NLP Analysis: Topic modeling (LDA, BERTopic), expertise inference, sentiment analysis
• Temporal Analysis: Change point detection, onboarding integration tracking, network evolution
• Multi-modal Fusion: Combines insights from multiple data sources
• Anomaly Detection: Detects isolation, overload, and temporal anomalies
• Intervention Framework: FINDING → HYPOTHESIS → INTERVENTION → MEASUREMENT workflow
• Ego Network Analysis: Personal network analysis for individuals
• Cross-Modal Validation: Validates insights across different data sources
• Team Stability Analysis: Time-Size Paradox analysis for team retention and stability
• Bonding/Bridging Analysis: Analyzes within-group (bonding) vs between-group (bridging) connections
• Reporting: HTML reports and static or interactive network views (Pyvis, etc.)
• Privacy-First: Built with privacy and ethics considerations

Installation

Dependencies are declared in pyproject.toml and pinned in uv.lock. Use uv for reproducible installs, or install with pip and extras as usual.

# Option A: uv (recommended — uses uv.lock)
uv sync --all-extras

# Option B: pip editable install with optional stacks
pip install -e ".[all]"

# Download spaCy language model (NLP extra)
python -m spacy download en_core_web_sm

Core-only install: uv sync or pip install -e .. After changing dependencies in pyproject.toml, refresh the lockfile with uv lock and commit uv.lock.

Quick Start

from orgnet.core import OrganizationalNetworkAnalyzer

# Initialize analyzer
analyzer = OrganizationalNetworkAnalyzer(config_path="config.yaml")

# Load data
analyzer.load_data()

# Build graph
graph = analyzer.build_graph()

# Run analysis
results = analyzer.analyze()

# Generate report
analyzer.generate_report(output_path="report.html")

Testing

Run the test suite:

pytest

For coverage report:

pytest --cov=orgnet --cov-report=html

See tests/README.md for more details.

Code Quality

Format code with black:

black orgnet/ tests/ example.py --line-length 100

Check code style with flake8:

flake8 orgnet/ tests/ example.py --max-line-length=100

Or use the Makefile:

make format   # Format code
make lint     # Check style
make test     # Run tests
make check    # Format, lint, and test

Project Structure

ONA/
├── orgnet/
│   ├── __init__.py
│   ├── core.py                 # Main analyzer class
│   ├── config.py               # Configuration management
│   ├── data/
│   │   ├── __init__.py
│   │   ├── ingestion.py        # Data ingestion layer
│   │   ├── models.py           # Data models
│   │   └── processors.py       # Data processing utilities
│   ├── graph/
│   │   ├── __init__.py
│   │   ├── builder.py          # Graph construction
│   │   ├── weights.py          # Edge weight calculations
│   │   └── temporal.py         # Temporal graph handling
│   ├── metrics/
│   │   ├── __init__.py
│   │   ├── centrality.py       # Centrality measures
│   │   ├── structural.py       # Structural holes, core-periphery
│   │   └── community.py        # Community detection
│   ├── ml/
│   │   ├── __init__.py
│   │   ├── gnn.py              # Graph Neural Networks
│   │   ├── embeddings.py       # Node2Vec, etc.
│   │   └── link_prediction.py  # Link prediction models
│   ├── nlp/
│   │   ├── __init__.py
│   │   ├── topics.py           # Topic modeling
│   │   └── expertise.py        # Expertise inference
│   ├── temporal/
│   │   ├── __init__.py
│   │   ├── change_detection.py # Change point detection
│   │   └── onboarding.py      # Onboarding analysis
│   ├── visualization/
│   │   ├── __init__.py
│   │   ├── network.py          # Network visualizations
│   │   └── dashboards.py       # Dashboard components
├── config.yaml                 # Configuration file
├── pyproject.toml              # Package metadata and optional dependency groups
├── uv.lock                     # Resolved dependency lockfile (uv)
└── README.md                   # This file

Configuration

Edit config.yaml to configure:

• Data source settings
• Graph construction parameters
• Analysis preferences
• Privacy settings

Privacy & Ethics

This tool is designed with privacy in mind:

• Aggregates data before storage
• Respects retention policies
• Focuses on patterns, not individual monitoring
• Configurable anonymization

Important: Always obtain proper consent and follow organizational policies before deploying.

License

MIT License

Contributing

Contributions welcome! Please read the contributing guidelines first.