Starting and running a successful graph project: Tips and advice from graph experts

Implementing a project with graph analytics and visualization can be challenging and complex. These projects often involve many key players, several different technologies, and multiple data sources, just to give some examples. Starting off on the right foot and making the right decisions along the way can minimize frustration and maximize the value your organization gets out of its graph project. 

During our Linkurious Days event in Paris, we gathered people passionate about graph technology in the same place for an entire day of meaningful collaboration and insights on what it takes to succeed with a graph project. A few of the experts who joined us for the day agreed to interviews to share their top tips and insights for graph project success.

A high-tech project also begins with the human factor. When starting a graph project, designate someone responsible for coordinating the team and all related tasks from beginning to end.

Thibault Kurc, who led several transformative graph projects at Credit Agricole and is now the founder and consultant at TKAI, emphasizes, “It's important to have this cooperation and it's also important that this person has multiple skills: business and legal skills, as well as technical skills and expertise in the field where we will derive value from the tool.”

It's crucial to manage the project in phases rather than attempting to complete all tasks at once. Initially, focus on essential features that will provide the most value for the company, and later, if beneficial, proceed to the optional features.

Another important step is putting the data into a graph for easy manipulation. Starting early makes the basics easier to handle. Vandy Berten, from SMALS, advises, “Think about what you are going to put in to begin with. You can use the Neo4j community edition, there are basic visualization tools that are not advanced but are good to start with and see where you can go.”

Consider seeking support for this phase, as Gwenhael Gauché, from GPIS, did: “Start building the database. Why not get support from consulting firms which helped us start and continue to support us in going further.”

A successful graph project depends on having all steps organized, with team roles well-defined, and understanding the available support and tools. However, there are additional tips to ensure great results.

Success lies in the methodology. Kurc elaborates, “The key to success for me is perfect cooperation between the business teams, technical teams, and of course the decision makers. It gives you the best chances of success for a graph project in a company.” Aligning teams with the same goals is crucial for achieving good results.

A well-integrated team is not enough without a good database. Gauché explains, “This database needs to be rigorous, solid. We need to be able to write everything the same way, which will allow us to go back to prior events and create correlations.”

A successful project also needs a wow factor—something that makes a significant difference in the final result and team’s experience. Using a good platform should facilitate this. Conduct a small trial to show users the project’s potential. Berten says, “We create that attention very early in the project, and so we can show the end user what data we will be able to use and what it looks like. Then very early on we can involve them in the process of developing the data model, the moving parts, and the expansions that happen.”

With this groundwork, you are ready to start your project. Having the right tools will also contribute to its success and enhance team engagement.

Starting a new project requires thorough research and a solid understanding of the fundamentals. Our article, "Charting a Course for Graph Project Success: Key Steps and Best Practices," provides extensive information, including an infographic detailing each step and offering additional advice. Check it out for more insights!