I am working on a credit research that compares customer profitability and default risk.

Since the customer profitability is confidential, normalization has been used in order to share the data and analysis.

The problem is that R² for the normalized data is different from the original data.

In other words, R² for Profit and Default Risk is not equal to R² of normalized Profit and Default Risk

To make sure that the data was not compromised in the normalization process, I plotted both and made sure that just the scale has changed.

Should Default Risk (which is a 0-100% number), be normalized as well, so the proportions (distance) remain the same?