Apply Differential Privacy to a HealthTech Analytics Dashboard
Overview
What this challenge is about.
Wrap the existing analytics layer with OpenDP (or Google's differential-privacy library). Implement epsilon-delta accounting: per-query Laplace noise for counts and sums, Gaussian mechanism for averages, and an adaptive composition tracker for repeated queries by the same user. Set an initial total budget of epsilon = 5 per dashboard user per quarter. For the top 12 product-manager queries (provided, anonymized), report the noise-induced error vs the raw answer. Deliver the wrapped analytics module, a privacy-budget design document covering epsilon-delta choices and the rationale, and a 6-page memo to the Chief Privacy Officer covering residual risks and the unsuitable queries you blocked.
The Brief
What you'll do, and what you'll demonstrate.
Retrofit a clinical-engagement analytics dashboard with differential privacy that preserves utility for 12 queries while bounding per-user privacy loss.
Earning criteria — what you'll demonstrate
- Apply Laplace and Gaussian DP mechanisms to real queries
- Design epsilon-delta privacy budgets that bound long-term loss
- Quantify utility loss vs noise tradeoffs honestly
- Communicate DP guarantees to a non-technical privacy leader
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Career mappings coming soon.