Using Python visual in Power BI for calling ML Azure rest API works in desktop version but not when published

Question

I created a dashboard in Power BI desktop. I have a trained model from ML Azure which I already deployed and has it's rest API. I need to call this rest API from the dashboard itself using measures I created (not from the query editor). I did it using the Python visual to send the input data and get back the output from the rest API and plotting the result (a number). This works perfectly in the desktop version. I need to publish this dashboard to share with other members of my organization but in the web version the script gives a runtime error. ![69200-capture.png][1] How to make it work? [1]: /api/attachments/69200-capture.png?platform=QnA

Answer 1

anonymous user There is a documented procedure to connect a Azure ML workspace on powerBI desktop and select the model and service before publishing this.
In this case it looks like you used python visual to consume your REST API but after publishing it the same script might be failing to even connect to the API. Not sure if any port needs to be opened up when published to web.

I think you can try the above procedure again with powerBI desktop using query editor and then try to publish the report to web.