View a markdown version of this page

Product Recommendations Powered by Neo4j - Product Recommendations Powered by Neo4j
Product recommendations diagramFurther readingDiagram history

Product Recommendations Powered by Neo4j

Publication date: September 20, 2022 (Diagram history)

With this architecture, you can deliver personalized product recommendations in near real time. The solution uses Neo4j Graph Database (GDB) and Graph Data Science (GDS) on AWS for storing and analyzing connected data, Amazon EMR for data processing, Amazon SageMaker AI for machine learning (ML), and Amazon Kinesis for streaming ingestion.

Product recommendations diagram

Reference architecture diagram showing how to deliver personalized product recommendations by using Neo4j, Amazon EMR, SageMaker AI, and Amazon Kinesis.

The following steps describe the data pipeline and recommendation engine for this architecture:

  1. Ingest customer orders, product data, and clickstream data through Amazon Redshift, Amazon RDS, Amazon EMR, Amazon Simple Storage Service, and Amazon Kinesis.

  2. Transform and enrich data by using AWS Glue, Apache Spark on Amazon EMR, AWS Lambda, or other tools.

  3. Load bulk and batch data to Neo4j by using the Neo4j Spark Connector for Amazon EMR or Neo4j APIs running in Lambda.

  4. Stream near real-time data from Apache Kafka on AWS through the Neo4j Kafka Connector to the Neo4j database.

  5. Flow clickstream data from Amazon Kinesis through the Neo4j Spark Connector running on Amazon EMR to the Neo4j database.

  6. Store, query, analyze, and manage highly connected data by using Neo4j GDB and GDS. Deploy Neo4j Aura as a fully managed service on AWS, or run Neo4j Enterprise on Amazon Elastic Compute Cloud instances.

  7. Create features on Neo4j GDB and GDS. Export graph embeddings to SageMaker AI for training ML models.

  8. Explore data and present findings by using AWS Quick Sight and Neo4j Bloom visualization tools.

  9. Access Neo4j by using official drivers from Lambda or any application. Use a low-code approach with GraphQL API and the GRANDStack framework through Amazon API Gateway.

  10. Run the recommendation services by using Neo4j GDB and GDS at scale in near real time. Apply this framework to build similar recommendation engines for additional use cases.

Further reading

For additional information, see the following resources:

Diagram history

To receive updates about this reference architecture diagram, subscribe to the RSS feed.

ChangeDescriptionDate

Initial publication

Reference architecture diagram first published.

September 20, 2022

RSS subscription requirement

To subscribe to RSS updates, you must have an RSS plugin enabled for the browser you are using.

⌂ 换网址