Amazon Neptune simplifies graph analytics and machine learning workflows with Python integration
You can now run graph analytics and machine learning tasks on graph data stored in Amazon Neptune using an open-source Python integration that simplifies data science and ML workflows. With this integration, you can read and write graph data stored in Neptune using Pandas DataFrames in any Python environment, such as a local Jupyter notebook instance, Amazon SageMaker Studio, AWS Lambda, or other compute resources. From there, you can run graph algorithms, such as PageRank and Connected Components, using open-source libraries like iGraph, NetworkX, and cuGraph.
You can now run graph analytics and machine learning tasks on graph data stored in Amazon Neptune using an open-source Python integration that simplifies data science and ML workflows. With this integration, you can read and write graph data stored in Neptune using Pandas DataFrames in any Python environment, such as a local Jupyter notebook instance, Amazon SageMaker Studio, AWS Lambda, or other compute resources. From there, you can run graph algorithms, such as PageRank and Connected Components, using open-source libraries like iGraph, NetworkX, and cuGraph.
Today’s launch helps customers to build and innovate faster by simplifying workflows to extract analytical insights for use cases such as knowledge graphs, fraud detection, entity resolution, and security posture management. For example, you can run a Connected Components algorithm on your Neptune data using NetworkX to identify strongly linked communities of users. You can then run PageRank to find the most influential users in each community and update these users with a “Most Influential” label in Neptune. You can also use Python libraries such as XGBoost to compute embeddings or make predictions on graph data, the SageMaker Python SDK to train and deploy machine learning models, or the Deep Graph Library, which is available today with Neptune ML.
To get started, you can use the AWS Management Console or AWS CLI to provision a Neptune notebook, which is hosted by SageMaker. To learn more, see the open source documentation and three new data science tutorials on analyzing graphs for fraud rings, synthetic identities, and transportation logistics optimization.
This integration is available in all regions where Amazon Neptune is available. There are no additional charges for using this integration. Customers only pay for the resources provisioned to run a Neptune cluster and a SageMaker notebook instance, where Neptune notebooks are hosted. For more information on pricing and regional availability, refer to the Amazon Neptune pricing page and the AWS Region Table.