Graph data science with Neo4j : learn how to use Neo4j 5 with Graph Data Science Library 2.0 and its Python driver for your project / Estelle Scifo.

Format:

Book

Author/Creator:

Scifo, Estelle, author.

Language:

English

Subjects (All):

Non-relational databases.

Graphic methods.

Physical Description:

1 online resource (289 pages)

Edition:

1st ed.

Place of Publication:

Birmingham, England : Packt Publishing, Limited, [2023]

Summary:

The ever-increasing need of graph representation among data scientists for modeling complex relationships and extracting contextual information is addressed by the latest version of Neo4j. This book shows you how to set up a graph machine learning pipeline using Neo4j 5, its Graph Data Science Library 2.0, and its Python client.

Contents:

Cover

Copyright

Contributors

Table of Contents

Preface

Part 1 - Creating Graph Data in Neo4j

Chapter 1: Introducing and Installing Neo4j

Technical requirements

What is a graph database?

Databases

Graph database

Finding or creating a graph database

A note about the graph dataset's format

Modeling your data as a graph

Neo4j in the graph databases landscape

Neo4j ecosystem

Setting up Neo4j

Downloading and starting Neo4j Desktop

Creating our first Neo4j database

Creating a database in the cloud - Neo4j Aura

Inserting data into Neo4j with Cypher, the Neo4j query language

Extracting data from Neo4j with Cypher pattern matching

Summary

Further reading

Exercises

Chapter 2: Importing Data into Neo4j to Build a Knowledge Graph

Importing CSV data into Neo4j with Cypher

Discovering the Netflix dataset

Defining the graph schema

Importing data

Introducing the APOC library to deal with JSON data

Browsing the dataset

Getting to know and installing the APOC plugin

Loading data

Dealing with temporal data

Discovering the Wikidata public knowledge graph

Data format

Query language - SPARQL

Enriching our graph with Wikidata information

Loading data into Neo4j for one person

Importing data for all people

Dealing with spatial data in Neo4j

Importing data in the cloud

Part 2 - Exploring and Characterizing Graph Data with Neo4j

Chapter 3: Characterizing a Graph Dataset

Characterizing a graph from its node and edge properties

Link direction

Link weight

Node type

Computing the graph degree distribution

Definition of a node's degree

Computing the node degree with Cypher

Visualizing the degree distribution with NeoDash.

Installing and using the Neo4j Python driver

Counting node labels and relationship types in Python

Building the degree distribution of a graph

Improved degree distribution

Learning about other characterizing metrics

Triangle count

Clustering coefficient

Chapter 4: Using Graph Algorithms to Characterize a Graph Dataset

Digging into the Neo4j GDS library

GDS content

Installing the GDS library with Neo4j Desktop

GDS project workflow

Projecting a graph for use by GDS

Native projections

Cypher projections

Computing a node's degree with GDS

stream mode

The YIELD keyword

write mode

mutate mode

Algorithm configuration

Other centrality metrics

Understanding a graph's structure by looking for communities

Number of components

Modularity and the Louvain algorithm

Chapter 5: Visualizing Graph Data

The complexity of graph data visualization

Physical networks

General case

Visualizing a small graph with networkx and matplotlib

Visualizing a graph with known coordinates

Visualizing a graph with unknown coordinates

Configuring object display

Discovering the Neo4j Bloom graph application

What is Bloom?

Bloom installation

Selecting data with Neo4j Bloom

Configuring the scene in Bloom

Visualizing large graphs with Gephi

Installing Gephi and its required plugin

Using APOC Extended to synchronize Neo4j and Gephi

Configuring the view in Gephi

Part 3 - Making Predictions on a Graph

Chapter 6: Building a Machine Learning Model with Graph Features

Introducing the GDS Python client

GDS Python principles

Input and output types.

Creating a projected graph from Python

Running GDS algorithms from Python and extracting data in a dataframe

Dropping the projected graph

Using features from graph algorithms in a scikit-learn pipeline

Machine learning tasks with graphs

Our task

Computing features

Extracting and visualizing data

Building the model

Exercise

Chapter 7: Automatically Extracting Features with Graph Embeddings for Machine Learning

Introducing graph embedding algorithms

Defining embeddings

Graph embedding classification

Using a transductive graph embedding algorithm

Understanding the Node2Vec algorithm

Using Node2Vec with GDS

Training an inductive embedding algorithm

Understanding GraphSAGE

Introducing the GDS model catalog

Training GraphSAGE with GDS

Computing new node representations

Chapter 8: Building a GDS Pipeline for Node Classification Model Training

The GDS pipelines

What is a pipeline?

Building and training a pipeline

Creating the pipeline and choosing the features

Setting the pipeline configuration

Training the pipeline

Making predictions

Computing the confusion matrix

Using embedding features

Choosing the graph embedding algorithm to use

Training using Node2Vec

Training using GraphSAGE

Chapter 9: Predicting Future Edges

Introducing the LP problem

LP examples

LP with the Netflix dataset

Framing an LP problem

LP features

Topological features

Features based on node properties

Building an LP pipeline with the GDS

Creating and configuring the pipeline

Pipeline training and testing

Further reading.

Chapter 10: Writing Your Custom Graph Algorithms with the Pregel API in Java

Introducing the Pregel API

GDS's features

The Pregel API

Implementing the PageRank algorithm

The PageRank algorithm

Simple Python implementation

Pregel Java implementation

Implementing the tolerance-stopping criteria

Testing our code

Test for the PageRank class

Test for the PageRankTol class

Using our algorithm from Cypher

Adding annotations

Building the JAR file

Updating the Neo4j configuration

Testing our procedure

Index

Other Books You May Enjoy.

Notes:

Includes index.

Description based on print version record.

ISBN:

9781804614907

1804614904

OCLC:

1369066002