Imputation

Imputation is filling in missing values with estimated ones so you can keep using the rest of a record. Simple methods replace a gap with the column's mean or most common value; smarter methods predict the missing value from the other columns.

Imputation is a convenience, not magic: the filled-in numbers are guesses, and pretending they are real makes your data look more certain than it is. Better practice is to flag which values were imputed and, for serious work, to use methods like multiple imputation that carry the added uncertainty through to your final estimates.