anonimise

Anonimise, or anonymisation, is the process of transforming personal data so that individuals cannot be identified, directly or indirectly, by any means reasonably likely to be used. The aim is to reduce privacy risks while preserving enough data utility for analysis, research, or reporting. When data are truly anonymised, they are no longer considered personal data under many privacy laws, though achieving true anonymisation can be challenging.

Anonimisation is distinct from pseudonymisation, where identifiers are replaced with substitutes but a separate key exists to re-identify individuals. Pseudonymised data remain subject to privacy rules because re-identification is still possible with additional information. Anonymisation also differs from de-identification, which may still allow some residual risk of linkage.

Common techniques include suppression (removing or hiding identifying fields), generalisation or aggregation (reducing precision, such as grouping ages), masking (obscuring values), and perturbation (altering data values). Advanced methods involve privacy-preserving analytics such as differential privacy and record-level transformations. Concepts like k-anonymity, l-diversity, and t-closeness describe properties that reduce re-identification risk by ensuring individuals cannot be uniquely distinguished within a group or by limiting attribute disclosure.

Implementation involves a risk assessment to estimate re-identification risk, selecting appropriate techniques, applying them, and validating residual risk. Limitations include that some data cannot be anonymised without rendering it useless for analysis, and residual risks persist, especially with external or background information. Privacy regulations (for example, GDPR) influence how anonymisation is defined and enforced, shaping its use in data sharing and research.