Amazon Neptune Analytics now Integrates with GraphStorm for Scalable Graph Machine Learning
Today, we’re announcing the integration of Amazon Neptune Analytics with GraphStorm, a scalable, open-source graph machine learning (ML) library built for enterprise-scale applications. This integration brings together Neptune’s high-performance graph analytics engine and GraphStorm’s flexible ML pipeline, making it easier for customers to build intelligent applications powered by graph-based insights.
With this launch, customers can train graph neural networks (GNNs) using GraphStorm and bring their learned representations—such as node embeddings, classifications, and link predictions—into Neptune Analytics. Once loaded, these enriched graphs can be queried interactively and analyzed using built-in algorithms like community detection or similarity search, enabling a powerful feedback loop between ML and human analysis. This integration supports a wide range of use cases, from detecting fraud and recommending content, to improving supply chain intelligence, understanding biological networks, or enhancing customer segmentation. GraphStorm simplifies model training with a high-level command-line interface (CLI) and supports advanced use cases via its Python API. Neptune Analytics, optimized for low-latency analysis of billion-scale graphs, allows developers and analysts to explore multi-hop relationships, analyze graph patterns, and perform real-time investigations.
By combining graph ML with fast, scalable analytics, Neptune and GraphStorm help teams move from raw relationships to real insights—whether they’re uncovering hidden patterns, ranking risks, or personalizing experiences. To learn more about using GraphStorm with Neptune Analytics, visit the blog post.