option a) drop: simple drop nulls if it makes sense option b)imputation: 1) numerical values (applies to ordinal categories): impute with mean, median, max, min, or even impute with 0, whichever makes more sense. Advanced numerical imputation: neural nets/knn 2) categorical (text/numerical): impute with most/least frequent value (or with "other") 2) string/objects: impute with most frequent/least frequent (or with "other")