Graph databases

1. Graph databases

In the final video of this chapter, we'll look at one more family of NoSQL databases; graph databases.

2. What are graph databases?

In the past handful of years, graph databases have exploded in popularity as machine learning and artificial intelligence use cases have become more sophisticated. By formal definition, graph databases are NoSQL data stores that persist data in a network of nodes and edges. In graph databases, each node represents an entity, and each edge represents a relationship between those entities. These nodes can have attributes, or properties, that further describe them. To interact with graph databases, a generic type of programming language called "graph query language" has been developed. Cypher, Gremlin, and GraphQL are more specific languages that have been developed and socialized to work with graph databases. Graph databases and query languages leverage graph algorithms to store and traverse this data, making reading and writing data stored in graph databases both performant and efficient. Have more questions? Well, let's take a closer look at graph databases with an example!

3. A closer look at graph databases

This image shows a graph database representing a social network. In this graph, there are four nodes and four edges. These nodes are individuals, and the edges between them represent each individual's relationship. This representation of information is simple and easy to understand. The information is stored intuitively and easily retrieved with a graph query language. If this information were to be stored tabularly, there would be an additional layer of abstraction upon the individuals and their relationships with each other, making patterns more difficult to recognize and may even increase computational overhead.

4. What are graph databases used for?

In the last example, we used graph databases to store social data. This is one of the most common use cases for graph databases. In addition to social networks, graph databases are widely used for recommendation engines, as well as fraud detection. In general, graph databases are great for studying and relationships, an essential workflow for many data professionals. In general, graph databases shine when the relationship between data points is to be studied, and these relationships are complex. Graph databases make it easier to traverse deeply interconnected data and are especially useful for data scientists and machine learning engineers looking to identify patterns and relationships within a dataset.

5. Graph database providers

As graph databases gain traction and users, the number of providers continues to grow. Neo4j is by far the most popular graph database due to its open-source nature and mature feature set. A number of cloud providers, including Azure and AWS, have graph database offerings that are popular at the enterprise level. ArangoDB, a relatively new graph database, has exploded in popularity and continues to evolve as it fights for a foothold in very a competitive market.

6. Let's practice!

Graph databases are awesome, and can help set you apart as a data professional. Time to refine those skills with a little bit of practice. Best of luck!

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.