Earn a 50% discount on the DP-600 certification exam by completing the Fabric 30 Days to Learn It challenge.
Hello,
I am new to Azure machine learning, but have been going through their Auto ML tutorials, and have trained and deployed a few models - in partiuclar time series forecasting models. However, when I try to access the models within Power BI Desktop to run predictions on tables, the models do not seem to be automativally detecting the columns as I expected, or have seen in some posts online
Example is below in screenshots:
I've updated the date column in my dataset to a "Date/Time" column as mentioned in tutorials, and it makes no difference in automatic column detection.
What I currently see in Power BI Desktop, it is leading errors in the output of the model prediction:
Below is what I expected to see based on the trained model and tutorials online - auto detected columns so the model can make predictions - below is a time series demand model as well from Azure Machine learning:
It feels like I am missing a very simple step, but it is hard to find full tutorials online for time series specific models that go from training and deploying, to connection in Power BI - specfically for time series models.
Any help would be greatly appreciated!
Hi @Deware80 ,
You may refer to this video tutorial :Azure Machine Learning with Power BI [BI to AI Acceleration]
Best Regards,
Eyelyn Qin
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
Thank you @v-eqin-msft for the response.
Unforeunately, whil the video is insightful - it still does provide an walkthrough for Auto ML time series models - and connecting them successfully to a Power BI dataset for future forecasts.
That's the step I seem to be struggling with, specififcally to Times Series models - Power Query is not picking up the columns automatically that were used in Azure ML.