Run report

The byproduct of model-dataset pairs

As we mentioned earlier, a project sets the scene for most of the error analysis. After all, error analysis is all about analyzing when, how, and why models fail, embracing the process of isolating, observing, and diagnosing erroneous ML predictions, thereby helping understand pockets of high and low performance of the model.

The run report is one of the most important components you’ll find on the project page. It is a byproduct of the link between a model and a dataset and serves as the entry door to deeply understanding your model.

The run report

The computation of the run report starts right after you upload a model-dataset pair to a project.

After the run report is generated, you will see a block with a preview, featuring some basic aggregate metrics, the most common error classes, among others. Click on Open report to go to the report page.

The report page has three parts: the metadata (shown on the right-hand side), the Error analysis panel (on top), and the data slice (on the bottom).

The metadata displays the model’s performance on the dataset, measured by aggregate metrics, and other information, such as date of creation, model version, and others.

What is fundamental to understand is the relationship between the Error analysis panel and the slice of data at the bottom. By playing with the error analysis panel, you can slice and dice your dataset. Thus, the data slice you see at the bottom is a consequence of your actions in the error analysis panel.

The error analysis panel and the data slice

Feel free to play a little bit with the error analysis panel, particularly on the Data distribution and Feature importance sections. Click on a few of the elements that appear on the error analysis panel. Can you see what happens at the data shown in the bottom?


Actionable insights

The first step in error analysis is usually identifying data slices where the model's performance is not where it should be. The error analysis panel provides an interface that can help you identify and isolate such slices, as we'll see in the next sections.

Now that you understand the relationship between the error analysis panel and the data shown at the bottom, let’s move on to the next part of the tutorial!