GuideData & AI
Anonymising data before sharing
By Inara Crest Engineering 1 min read
Removing names is not enough. Remove direct identifiers, reduce indirect ones and assess re-identification risk before sharing data.
Research and analysis often require sharing data with partners. Protecting the people in the data is both an ethical and a legal duty.
1. Remove direct identifiers
Names, ID numbers, phone numbers, email addresses and exact addresses.
2. Reduce indirect identifiers
Combinations such as age, location and occupation can identify individuals. Generalise them — age bands, larger areas — or remove them.
3. Watch small groups
Suppress or combine categories with very few people.
4. Assess risk
Consider who will receive the data and what else they could link it with.
5. Control access
Share through agreements that limit use and prohibit re-identification.
6. Document
Record what was changed, so analysts understand the data.
See our data services.
About the author
Inara Crest Engineering
Engineering team
Written by the Inara Crest Technologies engineering team, who design, install and support software, network, security, power and data centre systems for organisations in Kenya and East Africa.