Social Network Analysis

From the structure of real social networks to who spreads, and who stops, a contagion. Master in Computational Social Science, Universidad Carlos III de Madrid.

R igraph SIR MIT License
Epidemic threshold curve on the Copenhagen Bluetooth network.
Final epidemic size against the infection rate on the Copenhagen Bluetooth network. Below the critical rate the outbreak dies out; just above it, a sharp transition carries the contagion across most of the network.

The three deliverables

Final Work

Spreading on a Social Network

SIR dynamics with seeding, robustness, and quarantine experiments, replicated on a second, larger network.
Homework 2

Link Prediction

Topological heuristics against a random forest on the Bluetooth proximity network.
Homework 1

Descriptive Analysis

Degree distributions, community detection, and a configuration-model null comparison.
Two real networks. The Copenhagen Bluetooth proximity network (662 nodes, 9,251 edges) and, for a generality check, the arXiv GR-QC collaboration network (4,158 nodes, 13,422 edges). Targeting the most central people accelerates a contagion as seeds and suppresses it as blockers, and the findings hold across both networks.