1.3.3. Scaling and Standardizing data

When data is going to be used within a statistical model (like we will see in Section 3) or for machine learning (like we will see in Section 4) one often first applies a transformation. The most common transformations are:

• Min-Max scaling: One subtracts the minimum and then divides by the range to obtain data with minimum 0 and maximum 1.

• Standardizing (also called converting to a standard score or a Z-score): One subtracts the mean and then divides by the standard deviation to obtain data with mean 0 and standard deviation 1.

Our readings, from Wikipedia, describe these scalings and others.

• Feature scaling

• Standard score

Questions

• How would you apply Min-Max scaling to a column in a Table and add the result as a new column?

• How would you Standardize a column in a Table and add the result as a new column?