When is the right time to invest the time and money into trying a graph database solution at your organization?
We have a vested interest in this question. Naturally, we’d love for the answer to always be “right away” whenever a potential customer is considering graph solutions. However, that’s not a winning formula in the long run, for anyone involved.
You need the right solution for whatever business challenges you’re faced with. While graph technology is powerful and our solutions are incredibly scalable, they aren’t the right technology for every scenario.
At the May 2022 Graph + AI Summit, Afraz Jaffri, Director Analyst – Research and Advisory at Gartner, provided an insightful keynote on this topic. Titled “Understanding When Graph Analytics and AI Is Best for Your Business,” Afraz examined the question in depth. His keynote concluded with five indicators organizations can follow when considering if graph is the right solution. Those indicators, which we’ll get to momentarily, are the focus of this article.
Graph is Gaining Traction
While we’re concerned most with finding the right scenarios for organizations to bring in a graph solution, it’s important to note that organizations are increasingly turning to graph for an expanding array of use cases.
According to Afraz, “By 2025, models which can incorporate context, such as graphs, will replace 60% of models built on traditional data. We’re actually seeing this right now with some of the large tech companies, eCommerce, and social network giants incorporating graph-based features and using graphs as their data model.”
This expansion is likely, in part, due to the flexible nature of the technology. Later in the presentation, Afraz underscored graph’s power, stating “The essence that makes graphs powerful is the ability to capture multi-dimensional attributes in a single, flexible model and to make the relationships between the things as important as the things themselves.”
Making Graph the Right Choice
As promised, let’s take a look at the five indicators your organization can look for to determine if graph technology is the right choice.
Look at the name of the structure
Business projects often contain clues that are overlooked by the teams implementing them. If your organization is faced with a challenge that’s dealing with datasets and you consistently see labels for the data like network, tree, taxonomy, ancestry, or hierarchy, you should take note.
These labels can tell you that the data and the relationships between the data are equally important, a key sign that graph solutions are worth considering.
You have Big Data
You’ve heard the term and those connected with it. You’ve got Big Data (capital B, capital D). You’ve got data that meets the three “Vs” of volume, velocity, and variety. And you need to capitalize on it.
“Graph analytics,” said Afraz, “are uniquely suited to understanding structures and revealing patterns in data sets that are highly connected.” That’s inevitably true with Big Data.
As business use cases evolve, it’s also true with a new term – “small and wide data.” In a sense, this can be considered a subset of big data, high-quality data sets in a variety of formats where context is inherent, used to feed analytics projects and come to real business insights.
Your questions are “Path-y”
Are you looking to your data to discover an optimal or alternate route? This is a “path-y” scenario, one where you might need to model or predict courses of action. This can look different depending on the use case in question, but it all comes down to problem-solving and efficiency.
Are you trying to increase the efficiency of a supply chain? Are you faced with building a prediction model for a disease in a geographical area? These are “path-y” problems uniquely suited to graph.
Relationships are more important than facts
By its nature, graph is suited to working in situations where relationships are important. Scenarios exist where you’ll need to remove the facts, such as identifiers in cases where privacy is important due to regulations, but you’ll still need to analyze the relationships between those data points.
Scenarios like network attacks, identity theft, regulatory compliance, and financial fraud all fit this indicator.
You’re hitting the wall with artificial intelligence and machine learning
Artificial intelligence and machine learning projects are meant to increase efficiency, but what happens when you run into barriers to scaling those initiatives?
This is a good indication that a graph approach might be warranted. Not only will graph help you overcome these barriers, but you might also discover new solutions and improvements over existing models by incorporating graph into those projects.
Explore Graph for Your Connected Data Problems
Using these five indicators, you can determine for yourself whether a graph solution is the right fit for your organization. To explore these indicators and go further into the possibilities of graph for many use cases you might be working on, check out the full keynote presentation from Graph + AI Summit.
Here are some resources to help you learn more here about graph: