Anonymiserte

Anonymiserte, in English often rendered as anonymized, describes data from which personal identifiers have been removed or transformed so that individuals cannot be readily identified. In data processing, anonymisation aims to preserve enough data utility for analysis while reducing privacy risks.

Techniques used to anonymise data include removal of direct identifiers (names, addresses, social security numbers), masking or replacing identifiers, data generalization (for example, ages grouped into ranges), suppression of rare values, and data perturbation or aggregation. More formal approaches include k-anonymity, l-diversity and t-closeness, which seek to reduce re-identification risk by modifying data so that individuals are indistinguishable within groups. Differential privacy adds controlled randomness to analytic outputs to protect individual contributions.

Legal and ethical considerations vary by jurisdiction. Under the GDPR, truly anonymised data is not considered personal data and is not subject to the regulation’s requirements. If anonymisation is insufficient or re-identification remains possible, the data may still be treated as personal data. Organisations commonly perform risk assessments to determine whether data is effectively anonymised and implement governance for data minimisation and retention.

Applications and limitations: anonymiserte data are widely used in research, statistics, public health, and open data portals to enable insights without exposing individuals. However, anonymisation can reduce data granularity and analytic utility, particularly for rare subgroups or when data are linked with other sources. Ongoing evaluation of re-identification risks is essential to maintain privacy protections while preserving data utility.