Correlation plot: incorrect calculation

[Anonymous]

Hi,

I have a problem with the Correlation plot custom visual in Power BI, it displays the visual as it should but the numbers are all wrong. Even if I use a very simple Excel file with just 30 observations the numbers I get in Power BI differ completely from the ones I get in Excel. I tried using the corrplot packhae directly in R (the same component that the Power BI visual uses) and there it gives me the exact same numbers as in Excel, so the problem is not in the visual itself - but in how it uses the data in Power BI.

I have tried changing the datatype of the columns from decimal to whole number and using Don't summarize as the default aggregation to really make sure it uses the raw data - but I still get incorrect numbers.

Has anyone else experienced this, and solved it somehow?

Thanks!

[v-jiascu-msft]

Hi @Anonymous,

The root cause is that Power BI removes the duplicates. The visual just uses a small part of the whole dataset. So the workaround is adding an index for now. I have consulted the Product Team and will update here later.

Best Regards,

[Anonymous]

Ah, that makes sense - thanks a lot for your reply!

I'll use the workaround for now  🙂

Best regards,

Fredrik  

[v-jiascu-msft · Accepted solution]

Hi Fredrik,

Good news. The Product Team will improve this visual in the next version. They also provide a more beautiful workaround, using a textual index column instead of a numeric index column. Please refer to the snapshot below.

Best Regards,

[Cortana]

Is it updated yet? I am facing the same problem. Is the index parameter added yet?

[Anonymous]

Thanks a lot for the quick feedback! It's so nice to work with a tool that just keeps on improving all the time - good work!

Regards

Fredrik