Skip to main content

Resolve Project Walkthrough & Tutorial

Resolve is a powerful and extensive module in Fluree Sense.

We’ve provided detailed descriptions on how Resolve works and how to create and run a Resolve project. But it was always felt that a video walkthrough will provide a more immersive and interactive experience for new users, as well as explain the best practices when creating a Resolve project.

The below walkthrough is a series of 7 videos where we start with explaining the pre-requisites of running Resolve, the objective, and then move on to creation of the Project, mapping to entity attributes, and finally training the project to generate the Golden Records.

So stay tuned in and enjoy the following tutorials available as part of our online support documentation.

What We’re Trying to Achieve with Resolve​

As you will see in the video below, the end goal is to generate Golden Records output: a set of definitive, useful, merged, and mastered data records from multiple sources, which can serve as an unequivocal source of truth.

Entities and Why They Matter for Resolve​

One of the pre-requisites of running a Resolve project is having the relevant entities in place with pre-configured entity attributes satisfying some minimum criteria.

Datasets to Use in a Resolve Project​

Datasets are another pre-requisite of the Resolve project. Your datasets should be registered and ready before creating the project.

Creating a Resolve Project​

Let’s look at the actual workflow and best practices of creating a Resolve Project now, based on some sample Data and Entities.

Reviewing & Finalizing the Entity Mappings​

Finalizing the Entity Mappings is a key step in creating a Resolve Project. These mappings define how matching and merging will occur to generate consolidated Golden Records.

Run the Project & Observe the Home Screen​

Once the project is run, the summary and other useful information are displayed in the project Home Screen. Let’s analyze and review these.

Train Your Project Further (Supervised Training)​

Like any machine learning project, the Resolve project may require more round(s) of Supervised Training, where users work on “Training Tasks” to provide more feedback and improve the model and confidence.