Hello all,

I seem to be stuck in a school project. I have created a linear regression model in knime and I want to use it to report in powerBI. Can anyone help me in doing so? I have scanned through a lot of posts but cant seem to figure it out.Some people say to write to database and there are some posts regarding rest API. I am a new user of both powerBI and Knime and would appreciate the help immensely.
I basically want to take the predictor model i have made and output the results in powerBI. A great plus would be if i can use the model in powerBI to analyse new data coming in from the database. Please advise.

Hi @axa830 -

Unfortunately we don’t have any nodes that support a direct connection from KNIME to PowerBI. I believe it’s possible to use the nodes under REST Web Services (e.g. POST Request and PUT Request nodes) to communicate with PowerBI, but I don’t have any example workflows that demonstrate this.

Normally I would suggest that you output your model using the PMML Writer node for use with other software, but I gather that PowerBI doesn’t support PMML - I could be wrong about that, though. You could definitely output model predictions (using the Excel Writer node, for example) for additional visualization in Power BI, if that helps.

You say this is a school project, so I understand you may be constrained by project requirements, but is there a reason you can’t apply model results and do further visualization with additional KNIME nodes?

Hello Scott !

Thanks for your reply. I will look inti the rest service and the excel output.

The only inclination to use powerBI is to have a better visualisation of the data. Knime is awesome for analytics but i like the UI/UX of powerBI for the end report. Thank you !

Also my other question would be that if i do use the excel export is it possible to utilise the knime server so that the model is executed automatically as new data comes in and push the excel file to the cloud which can then be connected to powerBI fo have refreshed data.

Yes, that’s certainly doable. However, KNIME Server is our commercial product and requires a license. (Maybe this isn’t a problem if your school already has an academic license…)

Still, if you were willing to forego some of the automatic execution and scheduling capabilities, you could update your model manually using only KNIME Analytics Platform, and push the results to the cloud.