Azure automated machine learning time series forecasting models integration with Power BI RRS feed

  • Question

  • My organization uses Power BI, and we would like to forecast R&D expense in the future (time series forecasting). We think that Azure automated machine learning might be a good solution because there is no clear pattern or trend in the R&D expense over time, therefore we want to use many inputs (input variables) in the forecast to increase its accuracy. We understood that Azure automated machine learning can use many different input variables to create its time series forecasting models.

    I have little knowledge about how these things work, therefore want to ask some questions about their functionalities:

    1) Is it that after Azure automated machine learning creates the forecasting models, we will need to manually import these forecasting models into Power BI to see the forecasted results in Power BI? That means there is no way to automatically integrate Azure automated machine learning with Power BI so that we don't have to import the models to Power BI manually? 

    2) As we get more data in the future, do we need to train the models again with new data? Or is there a way to have new data going into Azure automated machine learning automatically to constantly improve the models?

    Our wish is to have everything done automatically if possible, which means new data automatically goes from its source to Azure automated machine learning to constantly improve forecasting models with new data, then these constantly-improved models are automatically integrated to Power BI to show constantly-improved forecasted results. Is it at all possible?

    Wednesday, March 4, 2020 3:11 PM

All replies