Transferium Player Intelligence

Web Application

This project is a football player replacement finder built on a Neo4j graph database and a weighted cosine similarity engine. Given any player from a dataset of 18,405 FC 26 entries, it surfaces the most statistically compatible replacements — filtered by position, adjusted for overall quality, and informed by league and regional context.

The graph stores 276,075 pre-computed similarity edges. Finding replacements at query time is a single traversal, with no recomputation or table scans.

Major tools utilised:              

A full write-up covering every architectural decision behind this project is available in the essays section.

Click on the image below to launch the live application.




Book a Session