My $5MM Differential Privacy Visualizations
summary
Project claims that visualizations unlock $5MM in funding and convince regulators and data scientists by showing real outputs and tradeoffs rather than abstract definitions. It uses a one-page interactive built on synthetic data to demonstrate noisy statistics end to end, a 15 by 15 rule threat demo that exposes catastrophic reidentification with real numbers, and load shapes and histograms to show usable accuracy. Two large-scale releases arrive with ε=4.72 and δ=5.06⋅10^−9, and stakeholders sign on.
AI-written summary · about
chapters (5)
- 0:00Introduction
- 3:15First visualization that won initial funding
- 5:43Regulator threat scenarios
- 8:32Data scientist visualization
- 10:38Design principles
related talks
10mCommunicating Differential Privacy Guarantees to Data Subjects
20mShare data without revealing personal information
17mNegotiating Privacy/Utility Trade-Offs under Differential Privacy
47mDifferential Privacy from a Statistical Perspective–Obtaining Valid Inferences...
32mYou've got trust issues, we've got solutions: Differential Privacy
42mDifferential Privacy
27mPrivacy-preserving statistical analyses
30mAccuracy First: Selecting a Differential Privacy Level for Accuracy-Constrained...