cancel
Showing results for 
Search instead for 
Did you mean: 
Reply
Greg_Deckler
Super User
Super User

Correlation

Technique to do correlation in DAX with thanks to @konstantinos for teaching me about VAR statement. This creates a correlation between forecasting table and wages tables:

 

  • In forecasting, create the following custom column:
CorrelateX2 = 
// Gregory J Deckler - 2/18/2016
// Create a new column that subtracts the Sales from the average of Sales
//
// First, calculate the average of Sales, making sure to remove any filters on forecasting table 
VAR AverageX2 = AVERAGEX(ALL(forecasting),[Sales])
// Return the Sales from the current row minus the average of all Sales
RETURN ( [Sales] - AverageX2)
  • In wages, create the following custom column:
CorrelateY2 = 
// Gregory J Deckler - 2/18/2016
// Create a new column that subtracts the Wages from the average of Wages
//
// First, calculate the average of Wages, making sure to remove any filters on wages table 
VAR AverageY2 = AVERAGEX(ALL(wages),[Wages])
// Return the Wages from the current row minus the average of all Wages
RETURN ([Wages] - AverageY2)

 

  • Create the following measure:
Correlation2 = 
// Gregory J Deckler - 2/18/2016
// This measure calculates the correlation between Sales and Wages using previously created custom columns
// [CorrelationX2] (Sales) and [CorrelationY2) (Wages)
// [CorrelationX2] and [CorrelationY2] calculate the difference of the current Sales or Wages from the average
//
// First, sum the product of the difference in averages for Sales squared
VAR SumCorrelateX2SumCorrelateX2 = SUMX(ALL(forecasting),[CorrelateX2]*[CorrelateX2])
// Next, sum the product of the difference in averages for Sales multiplied by the related difference in averages for Wages
VAR SumCorrelateX2SumCorrelateY2 = SUMX(ALL(forecasting),[CorrelateX] * RELATED(wages[CorrelateY]))
// Then, sum the product of the difference in averages for Wages squared
VAR SumCorrelateYCorrelateY = SUMX(ALL(forecasting),RELATED(wages[CorrelateY]) * RELATED(wages[CorrelateY]))
// Return the correlation calculated using the formula:
// Sum the product of the difference in averages for Sales multiplied by the related difference in averages for Wages
// divided by the square root of the product
// the sum of the product of the difference in averages for Sales squared multiplied by
// the sum of the product of the difference in averages for Wages squared
RETURN (
SumCorrelateX2SumCorrelateY2 / SQRT(SumCorrelateX2SumCorrelateX2*SumCorrelateYCorrelateY)
)

 

 


@ me in replies or I'll lose your thread!!!
Become an expert!: Enterprise DNA
External Tools: MSHGQM
YouTube Channel!: Microsoft Hates Greg
Latest book!:
Learn Power BI 2nd Edition
0 REPLIES 0

Helpful resources

Announcements
Microsoft Build 768x460.png

Microsoft Build is May 24-26. Have you registered yet?

Come together to explore latest innovations in code and application development—and gain insights from experts from around the world.

charticulator_carousel_with_text (1).png

Charticulator Design Challenge

Put your data visualization and design skills to the test! This exciting challenge is happening now through May 31st!

May UG Leader Call Carousel 768x460.png

What difference can a User Group make for you?

At the monthly call, connect with other leaders and find out how community makes your experience even better.

Power BI Release May 2022 768x460.png

Check it out!

Click here to read more about the May 2022 updates!