Advanced Network Science
Network Science
ITA / ENG
Context: Advanced Data Science
This repository collects the practical challenges and capstone projects developed for the Advanced Data Science course. The focus is on Network Science: applying graph theory and linear algebra to model, analyze, and visualize complex real-world systems.
Project 1: Madrid Train Bombing Network
Analyzing the terrorist network behind the 2004 Madrid attacks (M-11) to understand criminal organization structures.
- Similarity Analysis: Constructed an Adjacency Matrix weighted by trust, kinship, and co-participation. Used Cosine Similarity and Hierarchical Clustering to reconstruct the terrorist cells without prior knowledge.
- Resilience & Percolation: Simulated attacks on the network to test its robustness.
- Compared different node removal strategies: Random vs. Targeted (Degree, PageRank, Betweenness).
- Result: The network proved resilient to random failures but fragile to targeted removal of “hub” nodes (high Betweenness).
- Cohesive Blocks: Used k-connected components to identify the unbreakable core of the organization.
Project 2: Interdisciplinarity in Science
A Scientometric analysis of citation flows between scientific disciplines to measure “Autarchy” vs. “Interdisciplinarity”.
- Flow Normalization: Implemented a Configuration Model to calculate Expected Flows ($E$) vs. Observed Flows ($F$).
- Statistical Testing: Applied X-test and G-test to filter out random noise and visualize only statistically significant citation pathways.
- Entropy Measures:
- Shannon & Simpson: To measure diversity.
- Rao Quadratic Entropy: To measure true interdisciplinarity by accounting for the dissimilarity between cited disciplines (e.g., Physics citing Sociology is more interdisciplinary than Physics citing Math).
Project 3: NBA Assist Dynamics
(See dedicated project card) Applying centrality metrics to basketball tracking data to identify playmakers and ball-movement patterns.
Tech Stack
- Language: R
- Libraries:
igraph,ggraph,tidygraph,tidyverse,ggplot2, - Math: Linear Algebra (Eigenvalues, Spectral Radius), Statistical Mechanics (Entropy).
