Getting Started with Neo4j: Build Your First Knowledge Graph

Summary

In AI and data science, knowledge graphs are powerful tools for modeling complex relationships between entities. They enable intelligent querying, recommendation systems, and semantic search. Neo4j, an open-source graph database, is one of the most popular tools for building and managing knowledge graphs.

In this post, we’ll walk you through setting up Neo4j, configuring it for use as a knowledge graph, and manipulating the database using Python.

A knowledge graph consists of buckets of information (entities) with intelligent links between those buckets. Wikipedia, for example, can be seen as a massive knowledge graph connecting concepts, articles, and topics.

1. Download Neo4j

You can download Neo4j here:

Neo4j Download

Run the installation accepting the defaults.

2. Create a new Graph Database

Select “New -> Create Project” from the top left menu.

We will name it Knowledge Graph

3. Select a local DBMS and give it a password

Select Local DBMS from the drop down highlighted. You will need to enter a password for the database

Note: Make sure the graph database version is latest or at least > 5.17.0

Click on Create to create the DBMS.

4. Install the APOC plugin

This library has some useful functions for working with graph data.

5. Install the Graph Data Science Library

Install the Graph DataScience Plugin. It can be found in the plugins section of the explorer.

Note: In my case, I had to choose a lower version of the DBMS for compatibility with this plugin.

6. Modify your Neo4j.conf file

Open the configuration file directory

1. Backup the Configuration File: Always save a backup before making changes.
2. Enable Authentication: Uncomment the following line (remove # if present):

dbms.security.auth_enabled=true

3. Enable CSV Import: Ensure this line is present and set to true:

dbms.security.allow_csv_import_from_file_urls=true

4. Set Unrestricted Procedures:
   This is a development only setting. Don’t enable this on a production database.

dbms.security.procedures.unrestricted=jwt.security.*,apoc.*,genai.*

5. Allow Listed Procedures:

    dbms.security.procedures.allowlist=apoc.*,gds.*,genai.*

6. Copy neo4j-genai-plugin-*.jar from the products to the plugins directory.

• You can find the plugins directory using the menu you used to find the config. The products directory should be in your AppData folder `C:\Users[USERNAME]\AppData\Local\Neo4j..\products\neo4j-genai-plugin-5.20.0.jar1

After making these changes, restart Neo4j for the changes to take effect.

7. Start your database

8. Manipulating the database using python

from neo4j import GraphDatabase

class Neo4jExample:
    def __init__(self, uri, user, password):
        # Initialize the Neo4j driver
        self.driver = GraphDatabase.driver(uri, auth=(user, password))

    def close(self):
        # Close the driver connection
        self.driver.close()

    def create_entities_and_relationships(self):
        # Create nodes and relationships in a session
        try:
            with self.driver.session() as session:
                session.execute_write(self._create_nodes)
                session.execute_write(self._create_relationships)
        except Exception as e:
            print(f"An error occurred: {e}")

    @staticmethod
    def _create_nodes(tx):
        # Define queries to create nodes
        queries = [
            "CREATE (:Person {name: 'Alice', age: 30});",
            "CREATE (:Person {name: 'Bob', age: 25});",
            "CREATE (:Person {name: 'Charlie', age: 35});",
            "CREATE (:Company {name: 'TechCorp'});"
        ]
        for query in queries:
            tx.run(query)
        
    @staticmethod
    def _create_relationships(tx):
        # Define queries to create relationships
        queries = [
            "MATCH (a:Person {name: 'Alice'}), (b:Person {name: 'Bob'}) CREATE (a)-[:FRIEND]->(b);",
            "MATCH (b:Person {name: 'Bob'}), (c:Person {name: 'Charlie'}) CREATE (b)-[:FRIEND]->(c);",
            "MATCH (a:Person {name: 'Alice'}), (c:Company {name: 'TechCorp'}) CREATE (a)-[:WORKS_AT]->(c);"
        ]
        
        for query in queries:
            tx.run(query)

Creating the graph

export NEO4J_URI="bolt://localhost:7687"
export NEO4J_USERNAME="neo4j"
export NEO4J_PASSWORD="your_password"

import os
from dotenv import load_dotenv
load_dotenv()

uri = os.getenv("NEO4J_URI") 
user = os.getenv("NEO4J_USERNAME")
password = os.getenv("NEO4J_PASSWORD")

neo4j_example = Neo4jExample(uri, user, password)
neo4j_example.create_entities_and_relationships()
neo4j_example.close()
print("Graph database setup complete!")

You can then use the Database browser to view the data.

Click on the People button to see the graph.

Code examples

You can find the code here

NEo4j examples