Book Image

Graph Data Science with Neo4j

By : Estelle Scifo
5 (1)
Book Image

Graph Data Science with Neo4j

5 (1)
By: Estelle Scifo

Overview of this book

Neo4j, along with its Graph Data Science (GDS) library, is a complete solution to store, query, and analyze graph data. As graph databases are getting more popular among developers, data scientists are likely to face such databases in their career, making it an indispensable skill to work with graph algorithms for extracting context information and improving the overall model prediction performance. Data scientists working with Python will be able to put their knowledge to work with this practical guide to Neo4j and the GDS library that offers step-by-step explanations of essential concepts and practical instructions for implementing data science techniques on graph data using the latest Neo4j version 5 and its associated libraries. You’ll start by querying Neo4j with Cypher and learn how to characterize graph datasets. As you get the hang of running graph algorithms on graph data stored into Neo4j, you’ll understand the new and advanced capabilities of the GDS library that enable you to make predictions and write data science pipelines. Using the newly released GDSL Python driver, you’ll be able to integrate graph algorithms into your ML pipeline. By the end of this book, you’ll be able to take advantage of the relationships in your dataset to improve your current model and make other types of elaborate predictions.
Table of Contents (16 chapters)
1
Part 1 – Creating Graph Data in Neo4j
4
Part 2 – Exploring and Characterizing Graph Data with Neo4j
8
Part 3 – Making Predictions on a Graph

Making predictions

In order to make predictions, we are going to use the same projected graph that already contains the test nodes.

With this projected graph, and the model object returned by the pipeline training, we can now predict the class of new nodes:

predictions = model.predict_stream(
     projected_graph_object,
     targetNodeLabels=["Test", "Train"],
)

Note that the model object also exposes a predict_mutate function to store the results in the projected graph. This will be useful to us when dealing with embedding features in the last section of this chapter.

In the preceding code block, we include both the Test and Train nodes in order for the Louvain results to be computed properly, using the whole graph. We will filter out the predictions for the train nodes as we evaluate the model performances.

For instance, in order to evaluate our model, we can compute the confusion matrix using our...