How to Incorporate Azure ML Output to Power BI

In recent years, there has been an increase in data science and Machine Learning (ML) adoption across all industries, including within the Azure platform. Here, we’ll show you how to incorporate Azure ML output to Power BI.

Article originally published October 2020

More and more, companies are looking to expand traditional reports using data mining and predictive models to extract valuable insights, detect trends and forecast events. In response to these needs, Power BI provides several ways to integrate predictive models in reports and dashboards. Some of the options available are using R scripts and visuals, integrating data flows with Power BI AI capabilities, or using Azure ML output to Power BI (which I’ll be going over here).

In the previous blog about Demand Forecasting with Azure Machine Learning we did this integration to predict future customer orders based on a time series model trained with historical data.

The first step to integrate an Azure ML model with Power BI is to train said model using any of the options available in “Azure ML Resources”, and publish it accordingly. The model is then read from Power BI and integrated with a dataset. Afterwards, the model output is displayed in the Power BI dashboard.

Smartbridge is an Azure and Power BI Partner

An Example of Azure ML Output to Power BI

As shown in the image below, the model I have published for demand forecasting is available from the Power BI Azure ML feature. On the right side, it displays the option to map the features used for demand forecasting. In this case, region, item, season and brand need to be provided as actual inputs for the model.

Azure ML output to Power BI

After providing the corresponding input mapping for the predictive model, the forecasted metric is added to the dataset. Once the forecast is part of the dataset, it can be used in any report or dashboard.

Azure ML output to Power BI

One of the advantages of incorporating Azure ML output to Power BI is the seamless integration experience. If a model is retrained using a different dataset and republished, the new predictions will be available without any additional interaction needed.

A common limitation when integrating predictive analytics with a reporting tool is having to store predictions in a DB to be able to use them in a report. The possibility of skipping this step and connecting Azure models to Power BI directly simplifies the use of predictive analytics, and makes it more interactive.

Azure ML + Power BI = Seamless UX

Through the enhanced capabilities of Azure ML, organizations and end-users can enjoy the model-building experience with the other platforms supported by Microsoft, such as Power BI and R. By further exploring the depths of Azure’s ML features, users can break the bounds of traditional analytical reporting.

Looking for more on data & analytics?

Explore more insights and expertise at smartbridge.com/data

There’s more to explore at Smartbridge.com!

Sign up to be notified when we publish articles, news, videos and more!