Case Study: Artfinder achieves real-time, highly personalised art recommendations with Neo4j

Creating Powerful, Personalised Art Recommendations with Graph Databases

Artfinder is a London-based online marketplace connecting 600,000 users with 10,000 artists in 106 countries, where every user receives a personalised homepage. The company faced a “needle in a haystack” problem: hundreds of new artworks daily and users who struggle to describe their taste made it infeasible to deliver real-time, accurate recommendations with a relational database or opaque third‑party ML services.

Artfinder implemented a Neo4j graph database to power a real-time collaborative‑filtering recommendation engine that models rich relationships between users, artworks and artists. The solution delivered highly relevant, real‑time suggestions across hundreds of millions of relationships, improved discoverability and user engagement, and gave Artfinder a competitive lead in the online art market.

Artfinder

Jonas Almgren

MD