Philip Jama

Organizational Network Graph Visualization

Organizational Network Analysis

Meeting Patterns & Social Graph Mining

Project Overview

This project analyzes calendar event data to construct and visualize organizational social networks. Using graph-tool library and advanced algorithms including nested stochastic block models, edge betweenness centrality, and resistance distance calculations, it identifies collaboration patterns, community structures, and generates personalized connection recommendations.

Highlighted Organizational Network Graph
Viewed through a human lens, the people in this network are the nodes. As in biological systems like the C. elegans connectome, organizations tend to form small‑world structures: strong local clusters (teams) connected by a few well‑placed bridges. This modularity lets groups move quickly, while those cross‑team links keep information flowing efficiently across the company. Even as the organization grows, most people remain only a handful of introductions apart — preserving speed of communication without sacrificing focus.

The analysis processes meeting attendee data, applies Minkowski distance weighting based on meeting sizes (favoring smaller, more focused interactions), and uses hierarchical clustering to reveal organizational structure. Key techniques include:

  • Stochastic Block Modeling - Discovers hidden community structures within the organization
  • Resistance Distance - Calculates "effective resistance" between employees using electrical network theory
  • Edge Betweenness Centrality - Identifies key connectors and information brokers
  • Small-World Network Analysis - Measures organizational connectivity efficiency

The system generates personalized recommendations for each employee, suggesting colleagues they should connect with based on graph-theoretic proximity measures. Visualization includes hierarchical network layouts with highlighted connection paths and statistical analysis of collaboration patterns.

Project Details

Date:

Feb 2021

Dataset:

7,857 events, 227 calendars

Techniques:

Stochastic Block Models, Resistance Distance, Network Analysis

Tools:

Python, graph-tool, NumPy, Pandas, Matplotlib

Tags:

Graph Theory, Network Analysis, Machine Learning, Data Science