OutliersThe term data massage is informally used to indicate outliers in data were dropped because they were interfering with visual presentation or confirmation of a particular theory. As such, data massage has potential ethical, compliance and risk implications.
Data QualityAnother common type of data massage is to drop low quality records or to enhance their quality.
ExampleAn analysis of 1.2 million customers by location is failing because 4 records have no ZIP code. The data analyst manually drops them from the dataset and runs the report again.
|Overview: Data Massage|
Manually processing data to remove or fix entries that are breaking an automated processes or analysis.
Cleaning up small subsets of poor quality data that have questionable value.
Confirmatory Data AnalysisExploratory Data Analysis