Why are people moving to graph databases?

Dan McCreary

Why are people moving to graph databases?

I am interested in understanding the reasons that organizations are moving to graph databases.

I building a taxonomy of the use-cases for graph databases and I am interested in classifying the motivation for this switch.

I don't want to bias you too much, but here are some of my thoughts on the high-level structure of this taxonomy:
A Taxonomy of Graph Databases

Dan

Carol McGrath

RE: Why are people moving to graph databases?
(in response to Dan McCreary)
  • Flexible Schemas
  • High performance for complex transactions
  • High performance for deep analytics

William McKnight

RE: Why are people moving to graph databases?
(in response to Dan McCreary)

They're great when the workload calls for analysis of how things are connected and it yields great performance and ease of use when looking up those connections. They're also great for determining the importance of various items in and about the business. The built-in algorithms are very useful. And - the visualizations!

Phil Meredith

RE: Why are people moving to graph databases?
(in response to Dan McCreary)

Great article! I think you nailed it. We use a graph database for the key reasons you mentioned: flexible design, fast traversals, and the opportunity for unique analysis.

Regarding the classification issue, your project can die before it ever gets off the ground if you try to appease everyone. We implement tight rules along with a simple governance panel approach. I would be happy to show you how it all works. Send me a note: [login to unmask email]



Todd Hinton

RE: Why are people moving to graph databases?
(in response to Dan McCreary)

I think that one of the easiest ways to understand graph databases is to think about a social network like Facebook.  As a user on Facebook, you are connected to many different people (friends), and your connections are then connected to many other friends.  Think of these as connected nodes.  The friends you are connected to (nodes) then "like" things such as images, restaurants, etc.  This graph of connected nodes is what ties things together, so you will see things such as "your friend Bob likes the burger place on Main St.".  Here is a good article from neo4j on the Facebook use case. https://neo4j.com/news/facebook-graphsearch/

I have some other thoughts on this as it relates to where a graph database could come into play, and that is building a contact/customer graph.  RedPoint Global, the company I am part of, really pushes this message.  If we think about the 360 view of the customer and the many different touch points of a customer - physical, web browser, email, e-commerce, device data, social, transnational data - that could also be a use case for using a graph database.

Todd

Edited By:
Todd Hinton[All DATAVERSITY Members] @ Apr 05, 2019 - 04:35 PM (America/Mountain)

Thomas Deely, MSc in Tech Management, CFA

RE: Why are people moving to graph databases?
(in response to Todd Hinton)

Check out the article posted in the Columbia University Community Page and chat with me about our first ever conference on Knowledge Graphs for AI in the Enterprise at Columbia University in NY to learn more and discuss with 34 amazing industry speakers.

Edited By:
Laura Lucht[All DATAVERSITY Members] @ Apr 19, 2019 - 07:15 AM (America/Pacific)