In this workshop, you will build a product recommendation system by transforming relational data into a graph database. These systems need to find patterns in connected data—customers, orders, and products—and answer questions about relationships between them.
By the end of this workshop, you will build a product recommendation system that answers this question:
"What products do people like me buy, that I haven’t bought yet?"
Your workshop goal
You will import data from the Northwind dataset—a fictitious food products company with customers, orders, products, and categories.
91 customers
830 orders
77 products across 8 categories
Understanding the recommendation challenge
To answer this question, you need to:
Find what products I’ve purchased
Find customers with similar purchases ("people like me")
Find what products they bought
Exclude products I already own
Rank by popularity among similar customers
This is collaborative filtering—the algorithm behind recommendation systems provided by the world’s largest companies.
These questions are about relationships and connections, not just individual records.
What you’ll build
By the end of this workshop, you will have:
Complete graph model - Products, Customers, Orders, and Categories as nodes and relationships
Collaborative filtering algorithm - Find "people like me" based on purchase patterns
Recommendation query - 10 lines of Cypher compared to 38 lines of SQL
Performance understanding - How anchor nodes and traversal make queries fast
You will use the Import tool to design and load your graph model, then the Query tool to explore and query your data.
Summary
In this lesson, you learned about the workshop goal:
Goal - Build a recommendation system: "What products do people like me buy?"
The challenge - SQL requires 38 lines with 7 JOINs and complex subqueries
The solution - Cypher does it in 10 lines with direct relationship traversals
Performance - Graph traversal follows pointers in memory; deep traversals are more predictable and often much faster than equivalent multi-join SQL, especially as relationship depth increases
What you’ll build - Complete graph with Products, Customers, Orders, collaborative filtering algorithm
Workshop structure - 4 modules: Building Your Graph → Modeling Relationships → Many-to-Many → Building Recommendations
Dataset - Real Northwind data (91 customers, 830 orders, 77 products)
In the next module, you will learn graph fundamentals and import your first data.
