This repo contains notebooks demonstrating and visualising graph data science and ML using Neo4j and Graph Data Science with the Northwind and Cora datasets.

This demo compares classification using a tabular dataset vs using a graph dataset. Using Neo4j and the Graph Data Science libraries, we show that modelling the data using a graph allows us to achieve better classification accuracy!

The problem studied is classifying a dataset of academic papers into categories using the Cora dataset. Read the full tutorial here: "Comparing ML using tabular vs graph data models with the Cora dataset".

Here we show how querying a customer-product type of dataset may be easier with a graph data model vs tabular data model. This shows query commands using the Cypher language and their equivalents in SQL. We use the Northwind dataset of customers, products, suppliers and orders.

Rendered Quarto blog for Cora demo.