Building a real-time recommendation engine with Neo4j OSCON 2017 - PowerPoint PPT Presentation

Building a real-time recommendation engine with Neo4j OSCON 2017 William Lyon @lyonwj

William Lyon Developer Relations Engineer @neo4j will@neo4j.com @lyonwj lyonwj.com

Agenda • • • • Use cases • Recommender systems • Hands-on! • Graph data modeling • Cypher

Neo4j Graph Database • • • • •

Graph Data Model

Labeled Property Graph Model

Labeled Property Graph Model

The Graph

The Graph

The Graph

The Graph

The Graph

openCypher (queryLanguage)-[:FOR]->(graphs)

Cypher

Use Case: Personalized Recommendations

Personalized Real-Time Personalized Real-Time Personalized Promotions Recommendations Recommendations

Collaborative Filtering Content Based An algorithm that considers users An algorithm that considers Algorithm Types interactions with products, with the similarities between products and assumption that other users will categories of products. behave in similar ways. Product Product Category Customer Product Product Data-Model (Expressed as a graph) Product Product Category Customer

Polyglot Persistence Products Customers / Users Inventory Inventory Location Location Category Category Price Price Purchase Purchase Configurations Configurations Return Return Review Review View View In-store Purchases In-store Purchases Location Location RELATIONAL DB WIDE COLUMN DOCUMENT STORE RELATIONAL DB KEY VALUE STORE DOCUMENT STORE STORE Purchases Product Views User Review In-Store Shopping Cart Catalogue Purchase

Recommendations require an operational workload — it’s in the moment, real-time! Good for Analytics, BI, Map Reduce Non-Operational, Slow Data Lake Queries RELATIONAL DB WIDE COLUMN DOCUMENT STORE RELATIONAL DB KEY VALUE STORE DOCUMENT STORE STORE Purchases Product Views User Review In-Store Shopping Cart Catalogue Purchase

Apps and Systems Drivers: Java | JavaScript | Python | .Net | PHP | Go | Ruby Real-Time Queries Connector RELATIONAL DB WIDE COLUMN DOCUMENT STORE RELATIONAL DB KEY VALUE STORE DOCUMENT STORE STORE Purchases Product Views User Review In-Store Shopping Cart Catalogue Purchase

Graph-based recommendations

Collaborative Filtering

Collaborative Filtering

Collaborative Filtering

Collaborative Filtering

In Cypher

In Cypher

Content Filtering

Content Filtering

Content Filtering w/ Cypher

Content Filtering - Concept Hierarchy

Content Filtering - Concept Hierarchy

Content Filtering - Concept Hierarchy

Content Filtering - Concept Hierarchy w/ Cypher

Content Filtering - Concept Hierarchy w/ Cypher

Neo4j Sandbox

● → ○ ○ ● → ○ ○

Click “Neo4j Browser”

You should see this: Neo4j Browser ● Query workbench / visualization for Neo4j ● Interactive “guides” for our tutorial today ● Embed content, queries

(you)-[:HAVE]->(?) (?)<-[:ANSWERS]-(will)