Three terms that are often confused
- Masking is the technical measure: columns containing names, addresses or identifiers are left blank, replaced or altered.
- Pseudonymisation (Art. 4(5) GDPR) means: the data can no longer be attributed to a person without additional information that is kept separately. It remains personal data.
- Anonymisation means: attribution is no longer possible even with reasonable effort. Then the GDPR no longer applies. This is harder to achieve than it sounds.
Why masking still helps a lot
For most questions in manufacturing or controlling, nobody needs names. Scrap per line, contribution margin per item, lead time per work centre: this works without any link to individuals. If names never enter development, they cannot be shown there by mistake or sent to a model.
Two views of the same shape
For development, Onagero creates an analysis copy in which columns relating to individuals are masked. The tables have the same shape as the originals. This allows a tool to be built and tested on the copy.
When the tool is approved, it reads from a live view with the same shape, but only the columns and rows it is permitted to read. Whoever works in development does not see the live view. There is no way to unmask masked values in development.
Where masking reaches its limits
- Free text: In remarks and complaint texts, names appear in unexpected places. Such fields need separate handling and a check.
- Small groups: If a shift consists of two people, an analysis per shift reveals a lot despite masking.
- Combinations: Several harmless attributes together can make a person identifiable.
That is why we define the masking for every pilot field by field and document it so that your data protection team can approve it before the start.