Graph vs Relational: which is superior for AI workflows

Contributed Talk | Day 2 | 15:50:00 | 20 Minute Duration | GG-C

Did you attend? Rate this session

Graph vs Relational: which is superior for AI workflows

Contributed Talk | Day 2 | 15:50:00 | 20 Minute Duration | GG-C

Graph databases and algorithms are now one of the hottest technology areas.  Are they superior to the reigning Relational database approaches and destined to replace them?  Or just the latest fad that fails to live up to their hype and soon fades away?

Graph representations have gained traction because, in problem domains where the key is analyzing patterns of relationships in a network of entities, standard SQL can be a very clumsy tool, while the desired query in a graph database is often concise and fast.

So is the graph approach then superior to relational and should eventually replace it?  One defense of relational methods is that while there are situations where relational is cumbersome, there are likewise other kinds of information and queries where the graph is awkward and relational definitely wins.  The other retort is that graph technologies are unneeded, that relational databases really subsume graph databases — after all when you see a diagram of a relational model, are you not looking at a network of entities (tables) and relationships between them?

The answer is that both have their place.  This talk will provide an overview of graph technology and how to combine relational and graph approaches together into a hybrid data analysis workflow.