Integrating Azure Machine Learning Models into Power BI

While Power BI provides built-in automated visuals (Key Influencers, Decomposition Tree), complex enterprise predictive modeling (e.g., credit risk scoring, equipment failure prediction, medical diagnosis) is developed by data science teams in Azure Machine Learning (Azure ML).

Power BI enables direct integration with deployed Azure ML web services.

1. The Machine Learning Integration Pipeline

2. Invoking Azure ML Models in Power Query

1. 1
   In Power Query Editor, select the table containing input features (e.g., LoanApplications).
2. 2
   Go to the Home or Add Column ribbon tab and click AI Insights $\to$ Azure Machine Learning.
3. 3
   Power BI scans your Azure subscription and lists all published machine learning models you have permission to access.
4. 4
   Select the model (e.g., DefaultRiskPredictor).
5. 5
   Map table columns to model input parameters (e.g., map ApplicantIncome, CreditScore, DebtRatio).
6. 6
   Click OK. Power BI sends batch requests to the Azure ML endpoint and appends predicted outcome columns (ProbabilityOfDefault, RiskTier)!