
The Box Cox Power transformation creates a new dataset where the variables have been transformed according to the following transformation:

(Y^p -1)/p for p not equal to 0

log(Y) for p equal 0

Where p is controlled by a slider in a visualization specially designed to evaluate the asymmetry and linearity of the variables in the dataset after the transformation. When Y has values equal or below 0, a constant is added to the variable to make all the values in Y positive. This transformation is appropriate for continuous numeric asymmetric variables. Some interesting values of p are:
 
 2.0 - square - power of two
 1.0 - no transformation?
 0.5 - square root
 0.0 - logarithm
-1.0 - inverse
