Facebook Ego Network Analysis for Political Outreach Optimization

📌 Overview

This project analyzes Facebook ego networks to optimize political outreach by identifying high-influence users and communities. Using network science techniques, we:

• Detect algorithmic communities (Louvain method).
• Compare them to user-defined social circles.
• Simulate influence spread via threshold models.
• Prioritize nodes for cost-efficient advertising.

Dataset: Stanford SNAP Ego-Facebook

🔑 Key Features

• Network Analysis: Degree distribution, clustering coefficient, path length.
• Community Detection: Louvain algorithm vs. user-defined circles (Adjusted Rand Index).
• Centrality Metrics: Degree, betweenness, eigenvector centrality.
• Influence Simulation: Threshold model for targeted vs. random seeding.
• Visualizations: Network structure, community alignment, activation spread.

🛠️ Installation

1. Clone the repository:

   git clone https://github.com/yourusername/facebook-ego-network-analysis.git

2. Install dependencies:

   pip install networkx pandas matplotlib python-louvain scikit-learn tqdm

3. Download the dataset from Stanford SNAP and place it in /data/.

🚀 Usage

📊 Results

• Community Alignment: ARI = 0.144 (low overlap between algorithmic and user-defined groups).
• Top Central Nodes: Degree centrality outperformed other metrics.
• Influence Spread: Targeted seeding activated 92% of nodes vs. 7.8% for random.

📚 References

• Dataset: Stanford SNAP
• Libraries: NetworkX, pandas, matplotlib, python-louvain.