4 Data Cleaning and Preparation Free Online FlashCards

What does a missing value mean?

A missing value means data was not recorded or is unavailable; it is not equivalent to zero, an empty string, or “none.”

What should you do before counting missing values?

Identify and standardize markers such as blank cells, `N/A`, and `unknown` before counting missing values.

What is imputation?

Imputation fills a missing value with an estimate, such as a median for a skewed numeric column or the most common category for categorical data.

How should you decide whether records are duplicates?

Define a meaningful record key, such as a transaction ID or a combination of customer ID and order date. Repeated values alone do not prove that two records are duplicates.

What should guide the removal of suspected duplicates?

Inspect suspected duplicates before removing them, and document the rule used to decide which record to keep.

What must you check before parsing an ambiguous date?

Confirm whether dates use month-first or day-first order before parsing ambiguous values.

What should you verify before converting measurements?

Verify the units before converting measurements, such as centimeters and inches, so the conversion preserves the quantity’s meaning.

What does type validation check?

Type validation checks that values have the expected data types, such as dates that parse as dates and quantities that are numeric.

What does completeness validation check?

Completeness validation checks that required identifiers or other required fields are present.

What does range validation check?

Range validation checks whether values fall within plausible limits, such as whether an age or price is within an expected range.

What does allowed-values validation check?

Allowed-values validation checks that a value belongs to a defined set, such as `open`, `closed`, or `pending` for a status.

Does a failed validation check prove that a record is wrong?

A failed validation check is a reason to investigate, not proof that a record is wrong. A negative amount, for example, might represent a valid refund.