Interactive demo

See the methods work

Three live instruments — anonymise a table, generate synthetic data, or walk the full pipeline — driven by real sdcMicro / synthpop / riskutility results. Everything runs in your browser; no data leaves the page.

Protection level

Age 10-yr bands + k=5

age: 10-yr bands · region: exact · k=5

Re-identification risk 0.64% Safe

0%100%

Utility 84 %

Suppressed cells 0

min l-diversity 5

Released microdata sample k-anonymous

AGE	SEX	REGION	DIAGNOSIS sensitive — not a key	CLASS
90–99	F	Vaud	Migraine	25
80–89	F	Genève	Depression	33
0–9	M	Luzern	Diabetes	15
0–9	M	Zürich	COPD	73
20–29	F	Vaud	Migraine	203
30–39	M	Bern	Arrhythmia	372
40–49	F	Aargau	Hypertension	274
40–49	F	Genève	Asthma	239
50–59	F	Vaud	Migraine	312
60–69	M	Genève	Depression	182
60–69	F	Vaud	Hypertension	222
90–99	M	Aargau	Hypertension	17

Quasi-identifiers are generalized and the minimum number of individual cells are suppressed (sdcMicro kAnon) so each released record shares its key with others. Diagnosis is a sensitive attribute — never used as a key, but checked for l-diversity.

Real sdcMicro on a 25,000-record synthetic patient register. Metrics computed at build time.