Anonymization is hard. Just like cryptography, most people are not qualified to build their own.
Unlike cryptography, the research is far earlier-stage, and the pre-built code is virtually unavailable. That hasn’t stopped people from claiming certain datasets (like this ) are anonymized and (sadly) having them re-identified.
Full article: Deidentification versus anonymization