I've created a recipe that imputes missing numeric data using KNN. I did this on a test project using jupyter notebook in dataiku and would like to make it a global recipe so everyone can use it.
Here's where I'm stuck and am looking for documentation on how to proceed.
The data frame that we'll need to select and columns that need data imputed will be different per project. Does dataiku already have variables that I can use to select the dataframe and columns or do I need to create my own.
Also, any specific documentations that will help with global recipes?