A marker indicating that a data field has no value — distinct from zero or an empty string.
Null means unknown or absent. A zero means the value is known to be zero. An empty string means the field contains no text. These are three different things, and confusing them causes analysis errors.
Null values appear in datasets for many reasons: the information was not collected, the question did not apply, the data was not available at the time, or a system error occurred. Handling nulls correctly is one of the most important data quality tasks.
Example: A dataset of building permits has a column for demolition date. For permits that have not yet been demolished, that field is null — not zero, not a blank date, but genuinely absent. An analyst who treats null as zero would produce incorrect calculations.
Most analysis tools have specific functions for detecting and handling nulls, such as IS NULL in SQL or isna() in Python's pandas library.