Folks who want to make data-driven decisions naturally want data to make those decisions. In many cases, however, the data needs to be anonymized so the original user can be in no way identified, Humu Chief Privacy Officer Lea Kissner writes. There is a major challenge in building a data-slicing analytics system over anonymized data: applying the anonymization separately in different views of the dataset. This issue applies to different techniques, such as k-anonymity and differential privacy. Kissner covers these various pitfalls and the best ways to avoid them in this piece for Privacy Tech.
Full Story