Graph & Analytics
Transfer Network Analysis
Graph analysis of a synthetic transfer market: PageRank, corridors, fee outliers, flurries and boomerangs.
The Problem
Transfer data is a network. Spotting influential clubs and irregular deals needs graph algorithms, not just sorting.
How It Works
- Build a weighted club graph from 336 transfers across 26 clubs.
- Run PageRank (damping 0.85) with dangling-node redistribution.
- Flag anomalies: Tukey-IQR fee outliers, signing flurries and players returning to a former club.
Technically Interesting
Several independent analyses share one in-memory graph; the outlier fence (€7.85M) comes from the data’s own distribution.
Why It Matters
The same graph-and-statistics toolkit applies to supplier networks, customer relationships and fraud patterns.
Scope & honesty. Entirely synthetic clubs, players and fees.