Vanilla hot-deck

Hot-deck imputation is a simple method that replaces every missing value in a variable by the last observed value in this variable. It's very fast, as only one pass through the data is needed, but in its simplest form, hot-deck may sometimes break relations between the variables.

In this exercise, you will try it out on the tao dataset. You will hot-deck-impute missing values in the air temperature column air_temp and then draw a margin plot to analyze the relation between the imputed values with the sea surface temperature column sea_surface_temp. Let's see how it works!