sebastiandaschner blog

In the following content, I’ll explain how to build recommendations with graph databases, especially Neo4j. I’ll use an example that recommends coffee beans, based on their flavors and user rating.

As always, you can get the code on GitHub and feedback is highly welcome.

Part 1 — The setup

In this video, I’ll show the coffee example setup, how we use the Neo4j database from our code, and what we would like to create.

Our Neo4j database stores our coffee beans, its flavors, and our user rating, which are mapped into our Java code via Neo4j-OGM. We can use Cypher queries to access the data as well as the OGM functionalities to conveniently map the result set to our Java classes.

Have a look at the code, to check out the full example.

Part 2 — Creating recommendations

In this part, we’ll see how to recommend coffee beans by judging different criteria, and to build up our Cypher queries.

Part 3 — More recommendation queries

In this video, we’ll see more recommendations, what the users might like and what new flavors they should try out.

Part 4 — Running the examples yourself

In this part, we’ll see how you can run these examples yourself, with a locally running Neo4j Docker instance, or by using the cloud-hosted Neo4j Aura.

You find all the code in the following GitHub repo: https://github.com/sdaschner/favorite-coffee