Characteristics of Problems Best Suited for Graph Databases

·Jun 12, 2022 09:31 PM

every problem/data that i have encountered which is encoded in/via graphs can also be addressed with traditional relational databases. What are the characteristics of problems/data that are best suited for graphs ? Example: "use-cases involving measuring length of relationships when the records have hierarchical parent/child relationships". Concrete toy example "how long is the longest family tree in a large database of family trees" is more expensive to compute using relational databases. Do you know of more concrete examples for when graphs are advantageous ?

5 comments

· Sorted by Oldest

François S.
·
Any query that involves navigating through multiple relations and thus requires many joins will become computationally expensive in a relational database. However this is only one aspect. The flexibility of the schema, and the ability to model your domain according as it is naturally structured is also one of the big advantages fo knowledge graphs
👍1
Donny W.
·
When you expect that your understanding of your subject matter will evolve over time. When you want to stop worrying about whether you need a new table, or in which table to put a new column, or when to perform a db migration, or how to “inherit” properties from one table to another, or what you should name the new foreign key; when you want to solve problems in the language of your subject matter and not in the language of your rectangles du jour.
SMH Z.
·
Not sure whether this is what you're looking for, but post it anyway! For finding some hidden relationships between patients, symptoms and diseases, (like if they're from the same earning class, whether or not exercise, or smoke or ...), a study shows that the network features can be more representative than usual patient features. I think the advantage of graphs here is capturing the hidden structure or architecture within the data. Usually doctors go with age, gender and smoking status, but 5 more features added based on graph representation, and 2 of them are in top. So in this case the hidden (latent) relationships using Graphs are extracting more info than without graph.. Here's the article's title: A patient network-based machine learning model for disease prediction: The case of type 2 diabetes mellitus The PDF and the feature importance diagram are attached.
Shelley London
·
👍
Bob DuCharme
·
With relational databases, you must figure out your entire schema before you add any data. Well, you can evolve the schema after that, but it's a lot of trouble. The ability of graph schemas to evolve much more naturally lets your application evolve more naturally and also makes data integration much easier.

François S.
·
Any query that involves navigating through multiple relations and thus requires many joins will become computationally expensive in a relational database. However this is only one aspect. The flexibility of the schema, and the ability to model your domain according as it is naturally structured is also one of the big advantages fo knowledge graphs
👍1
Donny W.
·
When you expect that your understanding of your subject matter will evolve over time. When you want to stop worrying about whether you need a new table, or in which table to put a new column, or when to perform a db migration, or how to “inherit” properties from one table to another, or what you should name the new foreign key; when you want to solve problems in the language of your subject matter and not in the language of your rectangles du jour.
SMH Z.
·
Not sure whether this is what you're looking for, but post it anyway! For finding some hidden relationships between patients, symptoms and diseases, (like if they're from the same earning class, whether or not exercise, or smoke or ...), a study shows that the network features can be more representative than usual patient features. I think the advantage of graphs here is capturing the hidden structure or architecture within the data. Usually doctors go with age, gender and smoking status, but 5 more features added based on graph representation, and 2 of them are in top. So in this case the hidden (latent) relationships using Graphs are extracting more info than without graph.. Here's the article's title: A patient network-based machine learning model for disease prediction: The case of type 2 diabetes mellitus The PDF and the feature importance diagram are attached.
Shelley London
·
👍
Bob DuCharme
·
With relational databases, you must figure out your entire schema before you add any data. Well, you can evolve the schema after that, but it's a lot of trouble. The ability of graph schemas to evolve much more naturally lets your application evolve more naturally and also makes data integration much easier.