Amazon Web Services Inc.’s newest open-source project is Graph Notebook, a tool that enables data scientists to analyze and visualize information that their companies store in graph databases. 

Graph Notebook is based on software components that AWS originally developed for its Amazon Neptune managed graph database service. The cloud giant announced the project on Wednesday.

Graph databases are specialized databases that can store not only individual records, such as customer names and the dollar value of store purchases, but also relationships between those records, such as what customer made which purchase. This ability to capture relationships between data points is valuable for many use cases. As a result, graph databases are used to power recommendation engines, cybersecurity systems, scientific projects in areas such as healthcare and a wide range of other workloads.

Graph Notebook is designed to make working with information kept in graph databases easier. Even for savvy data scientists, manually finding patterns in large piles of records and data relationships can be a daunting task, which is why it’s often helpful to turn the raw information into a visualization. Graph Notebook makes that easier by providing the ability create visualizations from graph data using Jupyter Notebook.

Jupyter Notebook is a staple data science tool that’s extensively used in analytics projects because it allows users to run code against datasets and quickly see the results. With AWS’ Graph Notebook, data scientists can create a Jupyter Notebook query to extract the specific subset of the information in a graph database they want to work with, then visualize the information. The tool offers a choice of two query languages: Gremlin, which is widely supported by major graph databases, and the more niche SPARQL syntax.

Both Gremlin and SPARQL are specifically optimized for analyzing data stored in a graph format. Both are also supported by AWS’ Amazon Neptune managed graph databases. The cloud giant said that Graph Notebook can be used with Neptune, or deployed in another environment such as an EC2 instance or the user’s own computer.

Query results are visualized as a series of interconnected objects on a chart. A query that analyzes truck travel routes between a company’s warehouses, for example, might return a visualization in which each warehouse is represented by a circle and the distances between locations are visualized as connections between those circles.

AWS has made the code for Graph Notebook available on GitHub under an Apache 2.0 license. 

Image: AWS