What is a graph database?

A graph-based NoSQL database (commonly called a graph database, or GDB) organizes data as nodes and edges, and is designed to represent and query relationships between data points.

Unlike traditional relational databases that use structured tables, a graph database organizes data in a way that captures the relationships between data points. This structure mirrors some real-world networks, like people in a social network, products in a supply chain, or transactions in a financial fraud ring, making graph databases particularly powerful for helping analyze connections. Not all data takes this form, of course, but when relationships are central, a graph database makes it easier to analyze and uncover meaningful connections.

5:40

Spanner: The always-on, virtually unlimited scale database

Key takeaways

Graph databases are a category of NoSQL databases purpose-built to model and query relationships, using graph structures of nodes and edges to represent entities and their connections. Here’s a brief overview of their key features and benefits:

Relationship queries: Unlike relational databases that rely on tables and JOINs, graph databases prioritize relationship-centric queries, making them highly efficient for exploring connected data
Flexibility: Their flexible schema and compatibility with graph algorithms allow you to uncover patterns, optimize networks, and gain deeper insights from complex datasets
Variety of use cases: Common use cases include powering social media graphs, detecting fraud through anomalous connections, and optimizing logistics or route planning

What is a graph?

A graph is a data model that models relationships between entities. It consists of two key components:

Nodes: Represent entities such as people, products, locations, or events. These are your pieces of data.
Edges: Represent relationships between nodes, such as friendships on a social platform, connections between shoppers and purchased products, or supply chain links. Edges can also include properties, like timestamps or weights, adding context to the relationship.

How do graph databases work?

Think of graph databases as a large connect-the-dot puzzle. It stores your information as individual dots (nodes) and uses lines (edges) to directly show and store how those dots are related.

The flexibility of graph databases allows them to represent a variety of connections, from hierarchical structures (like family trees or organizational charts), to clustered networks (like e-commerce product recommendations), to the identification of influential nodes within social networks.

Specialized algorithms can enhance the insights graphs provide with:

Shortest path: Optimize routes in navigation and logistics
Community detection: Find groups of nodes that are more tightly connected to each other than to the rest of the network, helpful for social media segmentation or fraud detection
Web page ranking: Calculate the importance of a node based on the number and quality of links pointing to it

By leveraging algorithms like these, graph databases help transform complex relationships into actionable strategies.

Graph versus relational databases

While relational databases organize data into structured tables, graph databases focus on relationships. Here are some of the key differences:

Feature	Relational databases	Graph databases
Data structure	Use rows and columns in strict schemas with pre-defined properties. Adding a new relationship requires restructuring.	Model data as nodes and edges, allowing for flexible relationships without predefined schemas.
Query efficiency	Rely on JOIN operations to connect tables, which can become slow and complex as the number of relationships increases.	Traverse edges directly, making them faster and more intuitive for relationship-focused queries.
Query languages	Structured query language (SQL) and its derivatives.	Graph query language (GQL), Cypher, and Gremlin.
Use cases	Excel in structured, predictable environments such as financial systems or inventory management.	Ideal for applications where relationships are central, such as social networks, fraud detection, or route optimization.

Feature

Relational databases

Graph databases

Data structure

Use rows and columns in strict schemas with pre-defined properties. Adding a new relationship requires restructuring.

Model data as nodes and edges, allowing for flexible relationships without predefined schemas.

Query efficiency

Rely on JOIN operations to connect tables, which can become slow and complex as the number of relationships increases.

Traverse edges directly, making them faster and more intuitive for relationship-focused queries.

Query languages

Structured query language (SQL) and its derivatives.

Graph query language (GQL), Cypher, and Gremlin.

Use cases

Excel in structured, predictable environments such as financial systems or inventory management.

Ideal for applications where relationships are central, such as social networks, fraud detection, or route optimization.

Key use cases for graph databases

Graph databases typically do well in applications where understanding relationships between data points is essential. Here are some prominent use cases:

Graph databases power features like friend recommendations, influencer identification, and community detection. By analyzing the social graph and the connections between users, posts, and interactions, platforms can deliver personalized experiences and uncover key insights.

Fraud detection

In finance and e-commerce, graph databases can help detect fraudulent patterns by mapping transactions, accounts, and devices. They excel at uncovering hidden links, such as accounts sharing IP addresses or credit card details. When a node is highly connected to known fraudulent nodes, it raises suspicion.

Route optimization

Transportation and logistics companies rely on graph databases to optimize delivery routes. By analyzing nodes (locations) and edges (routes), they can minimize travel times, cut costs, and improve efficiency.

Product recommendation engines

For e-commerce retailers, graph databases can enhance personalized recommendations by connecting users to products they've interacted with (bought, viewed, rated) and to other similar products or users based on those interactions.

Major graph databases compared

There are multiple graph database vendors, with products that offer different features to suit specific graph use cases.

Along with dedicated graph database vendors, graph extensions for traditional databases are also available. Spanner Graph, for example, is a graph database product from Google Cloud, built upon the global-scale Spanner relational database. It combines strong consistency, horizontal scalability, and multi-region deployments.

Neo4j

Neo4j is a purpose-built graph database, available on Google Cloud, that offers high performance for complex queries like shortest path calculations and community detection. It uses a graph-optimized query language and is suitable for visualizing relationships for actionable insights.

AWS Neptune

AWS Neptune is a graph database service from Amazon Web Services. It supports popular graph models like property graphs and RDF graphs.

Enterprise Knowledge Graph

While not a database itself, Enterprise Knowledge Graph (EKG) is a solution from Google Cloud that uses graph principles to consolidate, standardize, and reconcile fragmented enterprise data from various sources. It helps create a unified, semantically rich graph model of an organization's knowledge, which can then be used to power advanced AI applications, contextual search, and a complete 360-degree view of entities like customers or products.

Spanner Graph

Google Cloud’s Spanner database combines relational and graph capabilities through Spanner Graph, offering global consistency, horizontal scalability, and the flexibility to manage graph and relational data in one unified environment, making it ideal for diverse and large-scale deployments.

You can adopt different types of graph databases depending on your needs. Dedicated solutions like Neo4j or AWS Neptune focus exclusively on graph-native operations, while multimodel databases such as Spanner Graph combine relational and graph models within one system, offering flexibility for diverse data requirements.

Get started with a Google Cloud Spanner free trial

FAQs about graph databases

What are "nodes" and "edges" in a graph database?

Nodes represent individual entities like people, products, or locations. Edges are the connections or relationships between those nodes, such as a friendship, a purchase, or a route between two locations.

Why would I choose a graph database over a relational database?

It may be beneficial to use a graph database when the relationships between data are as important as the data itself. A graph database is significantly faster and more intuitive for use cases like social networks, recommendation engines, and fraud detection, where you need to analyze complex connections.

Do I need both a relational and a graph database?

For many modern applications, you might need both. Think of your relational database (like PostgreSQL or MySQL) as the solid foundation of your data architecture. It's very reliable for storing the core facts of your business—your customers, products, and transactions—with strong data integrity.

If you also need to understand the complex, changing relationships between those facts, a graph database is useful. It's designed to answer questions about connections that are more cumbersome for a relational database, such as, "Which customers were influenced by this marketing campaign?"

The two databases can form a powerful partnership. Your relational database stores the 'what' (the customer, the product), while the graph database explores the 'how' (how that customer is connected to other customers and products).

Solve your business challenges with Google Cloud

New customers get $300 in free credits to spend on Google Cloud.

Benefits of graph databases

Graph databases are designed to handle highly connected data, offering advantages that can make them essential for relationship-focused applications. Here are some of the key benefits:

Faster queries for connected data

By directly traversing edges instead of relying on costly JOIN operations, graph databases deliver faster and more efficient performance for relationship-heavy queries, such as discovering the shortest path or detecting clusters.

Scalability for growing networks

Graph databases handle large and evolving datasets seamlessly, making them ideal for industries with dynamic data models, such as social media, finance, and telecommunications.

Flexibility in data structure

With schema-less designs, graph databases allow for easy additions or changes to nodes and edges without requiring significant reorganization. This flexibility supports scenarios where the nature of data relationships frequently change.

Specialized analytics and insights

Graph databases support advanced algorithms, like community detection and link analysis algorithms, to extract actionable insights from complex relationships. These capabilities are invaluable for uncovering hidden patterns and making data-driven decisions.

Intuitive relationship modeling

Graph databases use nodes and edges to mirror real-world relationships, making it easier to represent and analyze complex networks such as social interactions, supply chains, or recommendation systems.

Enhanced contextual awareness

By storing the meaning and type of relationships, not just the data, graph databases allow systems (particularly in AI) to gain a deeper understanding of the data, which can be critical for tasks like accurate contextual search and grounding AI models in verifiable facts.

Take the next step

Start building on Google Cloud with $300 in free credits and 20+ always free products.

Need help getting started?
Contact sales
Work with a trusted partner
Find a partner
Continue browsing
See all products