Getting Started with Neo4j Document Intelligence
Turn documents into a queryable knowledge graph with Document Intelligence
In this 45-minute course, you will learn
Neo4j Document Intelligence turns a collection of documents into a queryable knowledge graphA representation of real-world entities and their relationships, stored according to organizing principles, typically in a graph database. in AuraDBThe Aura product for transactional workloads., using AI to generate the graph model for you. In this course, you will take a set of synthetic doctors' notes from PDF files to a graph you can query with CypherNeo4j's implementation of GQL, the ISO standard query language for graph databases. It is declarative: you describe the pattern to find, and the database decides how to find it..
The Challenge of Unstructured Data
Understand why unstructured data is difficult to query with traditional tools and how Document Intelligence turns documents into a knowledge graph.
The Five-Step Pipeline
Learn the five-step pipeline that takes raw documents to a populated knowledge graph, including the lexical layer of chunks and embeddings.
The Document Intelligence UI and Agent
Navigate the Document Intelligence UI in the Aura console and work with the Document Intelligence Agent to generate and refine graph models.
Uploading, Modeling and Importing
Upload a document, generate and refine a graph model, and run an import into an Aura instance.
Verifying the Graph with Cypher
Restore a pre-built dataset and verify the resulting graph by running Cypher queries against it.
3 modules, 45 minutes.
Who this course is for
This course is for developers, data practitioners and analysts who have documents full of useful information and no easy way to query it. You do not need programming experience, an API key or any external account, because everything happens inside Neo4j AuraNeo4j's fully managed cloud service..
You do need an Aura account, a running AuraDBThe Aura product for transactional workloads. instance and generative AIModels that produce new content rather than classifying or scoring content that already exists. assistance enabled for your organization, which administrators can switch off. If you do not have an account yet, you can Create a Free Account. A free AuraDB instance is enough for this course, and Aura Fundamentals shows you how to create one. Familiarity with nodesA vertex in a graph. In a property graph it can carry labels and properties. and relationshipsA named, directed connection between two nodes. Every relationship has a type, a start node and an end node. helps, and Neo4j Fundamentals covers that ground.
What you will do
You will start with the problem: why information locked in documents is hard to reach with traditional query tools, and how Document Intelligence closes the gap with a five-step pipeline that takes raw files to a populated graph. You will then find your way around the Document Intelligence interface in the AuraNeo4j's fully managed cloud service. console, including the model canvas and the Document Intelligence Agent that helps you generate and refine models.
The hands-on work uses synthetic doctors' notes from a health claims scenario. You will upload a single note, describe your data domain in a context prompt, review the model the agent proposes and refine it in plain language. You will then run the import, watch the knowledge graphA representation of real-world entities and their relationships, stored according to organizing principles, typically in a graph database. appear in your Aura instanceA single Neo4j database running in Aura. and restore a larger backup of five notes. Finally, you will check the schema, constraints and indexes with CypherNeo4j's implementation of GQL, the ISO standard query language for graph databases. It is declarative: you describe the pattern to find, and the database decides how to find it. and run queries that link patients, doctors, diagnoses and procedures across the whole dataset — questions that would otherwise mean reading every note by hand.
Where to go next
To get more from the graph you have built, take Cypher Fundamentals to learn how to read and write queries yourself. Then try Building Agents in Neo4j Aura to create an agent that answers natural-language questions about your graph without writing application code.
Posts from the GraphAcademy blog.
From Neo4j Graph to Interactive Hive Plot: Modeling Graph Data for d3.js
The four modeling decisions behind an interactive d3.js hive plot of The Odyssey.
My five favorite cards in Aura Dashboards
When Dashboards first became available in Aura, I started to move my reporting away from custom tools. Here are the five cards I use the most, and a Cypher statement behind them.
EPIC: The Musical vs Nolan's The Odyssey - One Story, Visualised as a Graph
How does Christopher Nolan's film compare to EPIC: The Musical? I mapped every plot thread and character of Homer's poem in a graph database to find out.