Complex Networks Project — email-Eu-core analysis

Overview

• Project: Exploratory analysis of the email-Eu-core network from SNAP using code/cn-project.ipynb.
• Goal: Load the email network, attach department labels, compute structural statistics (components, diameter, degree distribution, clustering, assortativity, centrality, core decomposition) and test community detection algorithms.

Notebook

• Main notebook: code/cn-project.ipynb
  • Uses NetworkX and Matplotlib to reproduce the analyses and plots shown in the notebook.

Datasets

• Source: SNAP — https://snap.stanford.edu/data/email-Eu-core.html
• Files included:
  • dataset/email-Eu-core.txt: edge list (one edge per line, space-separated: source target). The file contains directed edges and a few self-loops.
  • dataset/email-Eu-core-department-labels.txt: mapping of node id to department id (one pair per line: node department). There are 42 departments in the ground-truth labels.

Requirements

• Python 3.8+ recommended
• Main Python packages: networkx, matplotlib, numpy, scipy

Quick setup

python -m venv .venv
.venv\Scripts\activate      # Windows
pip install --upgrade pip
pip install networkx matplotlib numpy scipy jupyter

Run

• Open the notebook in Jupyter or VS Code and run the cells in order. The notebook reads the datasets from dataset/ (relative path from code/ where the notebook lives).

Notes & observations (from the notebook)

• The graph has ~1,005 nodes and ~25,571 edges (directed), and is sparse.
• Degree distribution is heterogeneous but shows an exponential cutoff (not a pure scale-free power law).
• Clustering coefficient and modularity comparisons with null models and algorithms are included.
• Department labels provide a ground-truth partition (42 departments) used to evaluate community detection methods (Girvan-Newman, Louvain, greedy modularity).

Credits

• Dataset: SNAP (Stanford Large Network Dataset Collection)