Qlik Predict - Understand Explain-ability the "Why" Behind the Predictions

In this video, Mike Tarallo continues the "Do More with Qlik – Tips & Tricks Edition" series by diving into explainability data within Qlik Predict. Building on the previous episode where a simple predictive app was created, this video explains how SHAP (SHapley Additive exPlanations) values help you understand why a machine learning model makes specific predictions. Learn the difference between Prediction SHAP and Coordinate SHAP, how they contribute to transparency in predictive analytics, and how to add explainability insights to your Qlik dashboards. 🔍 Chapters: ⏱️ 00:05 – Introduction Mike introduces the topic and connects it to the previous video in the series. ⏱️ 00:12 – Recap: Predictive App with Prediction Data Set A quick review of how the prediction dataset was created from a Qlik ML experiment. ⏱️ 00:24 – Focus of This Video: Explainability Data An overview of what explainability data is and why it’s important. ⏱️ 00:35 – What Is SHAP? Introduction to SHAP values and their role in interpreting ML model predictions. ⏱️ 01:00 – How SHAP Values Work Explains how SHAP values are assigned to features based on their impact on predictions. ⏱️ 01:17 – SHAP and Feature Influence Discusses how SHAP helps identify which data features most influenced the model’s output. ⏱️ 01:24 – Explainability in Qlik Predict Details on how SHAP data is integrated in the Qlik Predict Compare and Analyze tabs. ⏱️ 01:36 – Two Types of SHAP Data Introduction to Prediction SHAP and Coordinate SHAP datasets. ⏱️ 01:44 – Prediction SHAP Explained Breakdown of how Prediction SHAP displays individual feature impact with positive or negative values. ⏱️ 02:12 – Coordinate SHAP Explained Explains the pivoted version of the SHAP dataset for easier visual analysis. ⏱️ 02:23 – Adding Explainability to a Dashboard Demonstration of how to add SHAP data to your predictive app and build visualizations.