Since a classic nomogram is typically 3 or more scales displayed on paper in such a way that a straight line connecting values on 2 of the scales will cross a predicted value on one of the other scales the proper question would be: can you place three (or more) sets of values in printed output to three (or more) fixed scales in such a relationship.
If ALL three scales are strictly linear with the same orientation on paper (i.e. vertical) I might attempt such if I was paid at an appropriate ad hoc job rate with a minimum overall. Since many times the curves of one or more of the elements of a nomogram are not linear and/or are at different orientations on paper this is very much not a trivial exercise.
What SAS will let you do easily with pretty much any of the regression procedures is to provide additional values of the independent variables and missing values for the dependent and then calculate estimated values for the dependent which would allow creating a look up table.
Just how many predictors are you talking about? (Each predictor would need its own representation)
How many of the predictors are categorical and how many continuous?
