NODES 2025 - A Sommelier's Guide to Recommendation Algorithms: Classical and Graph-Based Recommender Systems

Recommendation engines are all around us – on Netflix, Spotify, Amazon and many other platforms, they subtly shape what we watch, listen to, or buy. In fact, around 80% of what users watch on Netflix comes from recommendations, making these systems critical for user satisfaction and engagement. But building effective recommendation systems comes with challenges: massive datasets, complex user preferences, and the need for fast, accurate predictions. In this talk, Moritz will explore how to build a recommendation engine using the X-Wines dataset (https://github.com/rogerioxavier/X-Wines). He will start by presenting classical, non-graph-based approaches and then contrast them with a graph-based solution using Neo4j. The talk will walk through implementation details, key metrics for comparing different recommendation strategies, and a hands-on evaluation of both approaches. Finally, Moritz will highlight the strengths and limitations of each method and discuss when a graph-based system might provide a real advantage. Speaker: Moritz Wegener View Presentation: https://drive.google.com/file/d/142teMt5AdejnXk0d7r5wlyIbszKonyfm/view?usp=sharing Resources: Get Started with Aura - https://bit.ly/3LOLrjh Deployment Center - https://bit.ly/4jOelM3 Ground AI Systems and Agents with Neo4j - https://bit.ly/4oVsnyb #nodes2025 #neo4j #graphdatabase #graphrag #knowledgegraph