Correlation Plot 02-27-2019 11:35 AM

[Sharon]

Description

Correlation plots can be used to quickly find insights, showing what otherwise would require lots of iterative slicing and dicing of your data.

Download this sample to see how a correlation plot can inform marketing strategies in different market segments.

Prerequisites (The sample .pbix files will not work without these prerequites completed)

1. Install R Engine

Power BI Desktop does not include, deploy or install the R engine. To run R scripts in Power BI Desktop, you must separately install R on your local computer. You can download and install R for free from many locations, including the Revolution Open download page, and the CRAN Repository.

2. Install the required R packages.

Download the R script attached to this message and run it to install all required packages on your local machine. 

Required R packages:

corrplot

Tested on:

CRAN 3.3.1, MRO 3.3.0, powerbi.com

Legal Disclaimers:

Terms of Service and Third Party Programs.

[Katalo]

Hello!

I'm new to R visuals in Power BI. But I've installed Microsoft R Open 3.5.1 and the visual in the example. But when I insert values the visual gives ni this error:

Error Message:
R script error.
Loading required package: corrplot
Warning messages:
1: In library(package, lib.loc = lib.loc, character.only = TRUE, logical.return = TRUE, :
there is no package called 'corrplot'
2: In libraryRequireInstall("corrplot") :
*** The package: 'corrplot' was not installed ***
Error in corrplot(M, method = method, order = order, type = type, addrect = addrect, :
could not find function "corrplot"
Execution halted

What am I missing here?

[smpa01]

@Kataloprobably you figured it out by now. If not, I am not sure if the following steps solve your problem but this is my two cents. Initially, I faced a lot of trouble with R (got the similar message as you did). After a long trial and error and knocking on many doors I solved it by using the following

When I run R in PBI, the packages are accessed by PBI from C:\Users\xxxx\Library.

However, when I install the packages from R studio, the installed packages get saved in C:\Users\xxxx\Documents\R\win-library\3.5.

After installing a package, I copy that package

from

C:\Users\xxxx\Documents\R\win-library\3.5

to

C:\Users\xxxx\Library

Everything seems to work fine after that.

[JaouadZrhibi]

Hi Katalo,

I m really sorry, I am also a beginner in Power BI.

Hope you will find the solution.

Regards,

Jaouad

[boefraty]

Hi @JaouadZrhibi and @Katalo, 

Run

install.packages("corrplot")

from any R console. You can also download script provided with sample 
https://community.powerbi.com/oxcrx34285/attachments/oxcrx34285/RVisuals/12/16/ScriptInstallPackages...

Unfortunatly, if you use R Visual, power BI can not handle automatic package install. 

In case of R Custom Visual 
https://appsource.microsoft.com/en-us/product/power-bi-visuals/WA104380814?tab=Overview

It is installed automatically 

[Mike_Carlo]

The local installation of your R will need to include the "corrplot" package.  as boefraty stated below you can use the following code to install it. 

install.packages("corrplot")

You should run that in your local installation of R.  

Follow this tutorial to help you install thie package "corrplot" 

https://powerbi.tips/2016/09/using-r-visuals-in-power-bi/

[Anonymous]

For a while now, I am experiencing a problem with Correlation plot.

As I am using the same file sample with the same data and the same fields, but now importing the Correlation Plot visual from the visual market, I tried to replicate the same analysis, however, I always have the same error:

"Some columns are not numeric, or constant. Not enough input dimensions"

I have all the packages installed, everything, but always the same error, even if I type the code at the R script.

Any help with this?

Thank you in advance

[boefraty]

Hi @Anonymous, 

Make sure the columns are not aggregated. 

[ironryan77]

I love what you did here with this visual and it works perfectly in my PowerBI Desktop.  However, when I try to view this from Power BI intranet, I get this error.  I have tried to correct this by modifying the function and corrplot lines of code, with no success:

#verify if the column is numeric and non-constant
correctColumn <- function(someColumn) { !is.null(someColumn) && is.numeric(someColumn) && length(unique(someColumn)) > 1 }

corrplot(M, method=method, order=order, type=type, addrect=addrect, #na.label = "?",
na.rm=TRUE, 
          mar = defMar, tl.col = tl.col, tl.cex=tl.cex,  tl.srt=60, cl.ratio=0.2, cl.align="r",  
       number.digits=number.digits, number.cex=number.cex, addCoef.col=addCoef.col)

How can I fix this?

[boefraty]

Hi  ironryan77, 

Can you please share with me the PBIX, so I'd be able to reproduce this error?  

[ironryan77]

1. How does corrplot calculate correlation matrices?
2. Is this based on Pearson’s correlation coefficient formula? When I try plugging in my values to this formula, they do not match.
3. Does corrplot remove duplicate rows?
4. Does it handle aggregated columns?

[Mike_Carlo]

Here is the documentation on the corrplot function that was used. 

https://cran.r-project.org/web/packages/corrplot/corrplot.pdf

[JaouadZrhibi]

Hi,

After reading your nice posts, I need some help.

(Sorry, I m beginner on Power BI).

https://app.powerbi.com/view?r=eyJrIjoiNGNjYWEyMDItMmVhYS00OTVmLTgxM2ItYjFmMjYyODVmMDU5IiwidCI6IjBiO...

I would like to build graphs or indicators in order to see correlations between countries in import trade according 2 variables : weights in TON and volumes in USD.

I added this code but it doesn't work : 

require("corrplot")
library(corrplot)

M <- cor(dataset)

corrplot(M, method = "circle", tl.cex=0.6, tl.srt = 45, tl.col = "black", type= "upper", order="hclust")

If you have any suggestions, please advise.

Thank you in advance for your time and help,

Jaouad

[boefraty]

Hi @JaouadZrhibi, 

Your code is almost perfect just di right-click on your data clumns and select  "Don't summarize" 

[ironryan77]

Thanks, but I think I got this working now!

[Mike_Carlo]

This is a great R visual.  I have taken some time to build a complete tutorial on how to 1. Install R, 2. add R code to PowerBI, 3. Manipulate the code to change the visual.  You can follow along and learn how to build this here (www.PowerBI.Tips).

Enjoy,

Mike

[boefraty]

Great tutorial ! Here is another useful resource: https://powerbi.microsoft.com/en-us/documentation/powerbi-desktop-r-powered-custom-visuals/#troubles...