Problem
When JMP imports data, e.g., a CSV or an XLSX file, JMP guesses the appropriate Data Type and Modeling Type for each Column. This guess is helpful but is not (cannot be) perfect for every scenario. When a large table, e.g., one with hundreds of Columns, is imported, subtle imperfection of the guessing may lead to a lot of repetitive manual work afterwords. Specifically, JMP treats all Numeric Columns as Continuous.
Suggestion
I believe many users will benefit if some simple rule-based branching is added here. A specific example of such rules can be adding an option to treat Columns which
- contains only integers
- contains only a small number, e.g, 10 or less, of unique values
as Nominal.
I believe this applies to all recent versions, e.g., 14, 15, 16, and 17.