There's really three different options.
1. Normalize -> Select from Input Data
This is the basic setup. Select the numeric fields you like to normalize (by default it will select all numeric fields). This will produce your original dataset with all of the features normalized, as well as a parameters table that has the mean and standard deviation information.
2. Normalize -> Use Parameters Table (requires Parameters table)
This will normalize a table based off the parameters table from another dataset. If you want to run a model on new data (dataset B), but normalize your dataset based off of the dataset used to create the statistical model (dataset A), you can use the parameters table generated for that original dataset (A) to normalize your new dataset (B) from.
3. Denormalize (requires Parameters table)
This will take the normalized dataset and return your original data if you pass it the parameters table that was used to create it. The reverse of step 1 or step 2.
Note: Because of the way metadata flows through Alteryx, and since the data is transposed and cross-tabbed in the process, this macro may give you an error before you run it, even though it will actually work just fine. I'm not sure how to fix that, I think it's a bug with metadata within Alteryx itself. Please let me know if you run into any situations where the macro itself actually doesn't work (I haven't found any situations where the error wasn't a red-herring at this point).
