Customizing Prediction Studio

Hi! I'm Kevin, a Data Scientist who's new to Pega. I've been assigned to a Customer Decision Hub implementation project. I'm going to reach out to Bella, the Lead System Architect on the project, to get some recommendations on customizing Prediction Studio.

Hi Kevin, thrilled to have you on board!

Let's kick off by configuring the notifications that keep the data scientist team in the loop when attention is needed. Bella, can you tell me more about the notifications generated by Prediction Studio?

Well, Prediction Studio generates a variety of notifications that offer insights into the performance of your predictions, adaptive models, and predictive models. You can access these notifications by clicking on the Notifications icon in the header of Prediction Studio. These notifications are valuable for monitoring your models, updating their configuration if necessary, and enhancing the predictions that drive Customer Decision Hub decisions. If you want to modify notification settings, you can navigate to the Monitor & notification settings section in Prediction Studio.

Got it. So, just to clarify, I have to keep logging into Prediction Studio to read the notifications, right?

No, you can also receive daily email summaries that highlight high-impact notifications. So, you don't need to access Prediction Studio unless the notifications prompt you to do so.

That's great! I can stay informed without constantly logging in.

Now, I noticed there are different notification categories. In your opinion, which notifications in the Performance category are the most important?

Some of the crucial notifications in this category include sudden drops in predictive performance, models with low performance, and changes in lift or accuracy. The additional notifications you enable will depend on the specific use case.

I can imagine a flood of notifications coming our way. Is there any way to limit the number of notifications to avoid information overload?

Certain notifications have customizable thresholds that you can adjust according to your preferences. This allows you to define when you want to receive specific types of notifications. By customizing the thresholds, you can receive notifications at a frequency that suits your needs.

So, just to summarize, when implementing a new Customer Decision Hub project, the data scientist team configures the notifications that alert them to important issues impacting the models and predictions in the system.

Absolutely. These notifications help the team promptly address critical matters.

Thanks a lot, Bella! That was incredibly insightful!