Privacy and security in analytics
You minimise analysis data, apply masking, pseudonymisation, k-anonymity and differential privacy, measure re-identification risk and make access auditable.
Level 401 · 16 lessons · 12 practice labs
Lessons
Risk
- Privacy risk in analytics: why aggregated data can still be dangerous
- Re-identification attacks
- Data minimisation for analysis
- Quick task: find the quasi-identifiers
Techniques
- Masking and tokenisation
- Keyed pseudonyms: HMAC and key management
- k-anonymity and generalisation
- Differential privacy: safe counts with noise
- Quick task: anonymise HR data for analysis
Access and audit
- Column- and row-level access
- Encryption for analysis files
- Query audit logs
- Retention and deletion for analysis copies
- Quick task: an access review of the analysis environment
Sharing
- Synthetic data: when it helps
- Capstone: prepare a privacy-preserving data set for sharing
Practice labs
- Find the quasi-identifiers: Mark the columns in a survey data set that are harmless alone but identify a person together and write down why.
- A re-identification attack: Link an “anonymous” health data set to a public member list on birth year, postcode and sex and show how many people remain unique.
- Data minimisation for analysis: Decide which columns and level of detail a churn analysis question really needs and prove the reduced data gives the same result.
- Mask card and phone numbers: Mask the card and phone numbers in a call-centre export so only the last four digits remain.
- Keyed pseudonyms with HMAC: Turn customer IDs into pseudonyms with HMAC and show they cannot be reversed by a dictionary attack without the key.
- Refresh pseudonyms on key rotation: Rotate the pseudonym key, produce the old-to-new mapping table securely and retire the old key.
- Measure k: Find the smallest group size by quasi-identifier combination in an HR table and report its k value.
- Reach k=5 by generalisation: Generalise age into bands and postcode into regions to reach k=5 while suppressing as few rows as possible.
- Safe release with noisy counts: Add Laplace noise to sick-leave counts per department and tabulate the accuracy-privacy trade-off at different epsilon values.
- A row-level access view: Build a view that lets regional analysts see only their own region's rows and test it with two different users.
- Too much curiosity in the query audit log: Find searches in the query log that target a single person without a business reason and report them per analyst.
- Clean up forgotten analysis copies: Find copies with personal data past their retention period in a shared analysis folder and remove them with a deletion record.
About this course
You minimise analysis data, apply masking, pseudonymisation, k-anonymity and differential privacy, measure re-identification risk and make access auditable.
What you will learn
- Risk
- Techniques
- Access and audit
- Sharing
Prerequisites
You should have practised the topics of the same path up to level 301.
Lab environment
Lessons run in your own Kali Linux terminal that opens in the browser. Nothing to install; the lab starts clean in every lesson.
Frequently asked questions
- Is this course right for my level?
- The course is at level 401; it assumes you know the topics of the lower-level courses.
- Is there anything to install?
- No. Every lesson runs in your own lab environment that opens in the browser.
- How do lessons work?
- Each lesson has 8 to 10 steps and a short quiz. Early steps show the command to type, later steps give less help; if two attempts fail, the solution opens.
- Can I try it for free?
- Yes. The first three lessons of every course are free.