Knowledge Graph Best Practices | Selecting Your First Use Case

Posted by Sam Chance on Sep 14, 2021 8:00:00 AM
Find me on:

This is the third post in the Knowledge Graph Best Practices series. Follow along with the series for all the best tips and tricks to standing up your first knowledge graph project.

Seeing is believing and nothing breeds success like success. So the next thing you should do is develop a demonstration. The demonstration should concretely illustrate what a knowledge graph is, and how it helps users and decision makers. 

 

Nobody cares more about their own problems than themselves, so the data you use should be relevant to your audience. As a vendor, we have developed many demonstrations! But none are more effective than those that illustrate the capabilities using the audiences’ data. For example, we developed a Maritime Common Operating Picture (COP) demo using real world data from a geopolitically important region. We felt the power of this demo was that many audiences understand the domain, the need, and the data.

This next idea may seem contradictory, but it’s necessary. Identify a use case you know will resonate with your audience. Then use data from sources that support that use case. But consider ways the knowledge graph developed from these sources can be reused across other applications. Knowing what you know about “data centricity,” construct your knowledge graph in the context of a product. Represent to the audience the concepts and relationships, not the physical aspects. At a minimum, the concept-oriented view will show that you are doing something different. In other words, a user doesn’t necessarily care about “how” the data is delivered to them, but they will care about how it is represented (i.e., modeled) to them.

We have found knowledge graph technology is an easier sale to technical audiences, especially visionaries and architects. They more readily grasp the differences and can see the benefits. Nevertheless, data searching and preparation are so ingrained in the collective mindset that technical people often conclude that it’s just not necessary. But when they realize how it makes their lives easier, they lean forward to pay attention.

As an anecdote, I once worked on a project wherein the customer tasked us to survey the users of an existing system and ask them what they would like in the new system. I asked an analyst, “what if the system automatically connected and presented data from many sources for you?” The answer, which was revealing, was, “That’s my job!” At that moment, I realized we had a serious challenge.

So, what did we do? We created a knowledge graph from several sources, configured a dashboard with different views that surfaced data from many sources in a single pane of glass. Mind you, it was not a simple aggregation, or roll up, of multi-source data. It was a fully integrated and articulated knowledge graph. I showed the audience how they could configure different views based on their use cases or questions. From that point, they were hooked!

What it meant for them was that they no longer had to fetch data from many sources and then create complex spreadsheets of aggregated data, which was time consuming and error prone. They could simply point and click to filter data based on their questions. This was a game changer.

Your demo needs to show these capabilities and it needs to show them using operationally relevant data.

Some readers may already be putting up mental barriers — objections. in your mind. We address common barriers in this blog post.

What is stopping you from selecting your first use case?

Knowledge Graph Best Practice Series

  1. Building Momentum | Educating others about knowledge graph and getting support for the project
  2. Selecting Your Fist Use Case | Set up for success
  3. Assembling the Team | Required roles and skills
  4. Preparation | What to do before you begin to avoid delays (coming soon)

 

Tags: Analytics, Knowledge Graph

Subscribe to the Smart Data Blog!

Comment on this Blogpost!