Prior to releasing a dataset, a data custodian considers the intended uses of the dataset and determines if it is necessary to release personally identifiable information. If the personally identifiable information is not necessary, the information can be removed using de-identification techniques.