A Privacy Analysis of Cross-device Tracking
summary
Describes a measurement method that detects cross-device linking, a dataset from 126 users (107 with both mobile and desktop histories), and an algorithm using IP overlap, shared URLs and app-to-site mappings that achieves up to F1 = 0.91 for matching mobile to desktop in this set. Reports ~9,732 desktop and ~3,000 mobile trackers with 2,571 overlapping, Google reaches ~50% of users, Facebook ~20%, and finds 17 of 40 sampled companies omit cross-device disclosure in privacy policies.
AI-written summary · about
chapters (10)
- 0:00Introduction to tracking
- 4:23Methodology and experiments
- 6:45Data flow analysis
- 10:04Device matching algorithm
- 14:15Performance metrics
- 15:40Insights from user data
- 19:11Landscape of trackers
- 21:44Privacy policy compliance
- 23:38Future work and conclusion
- 25:04Q&A session
related talks
49mWeb Side Story: A Privacy Odyssey through Tracking and Fingerprinting
27mTracing Information Flows Between Ad Exchanges Using Retargeted Ads
15mA Closer Look: Evaluating Location Privacy Empirically
34mCovert web-to-app tracking via localhost on Android
49mBeyond Cookies: The Unseen Privacy Risks of Web Analytics
27mTrack all users! How to locate & track users and bypass all those nagging privacy...
06mReplication: Why We Still Can't Browse in Peace On the Uniqueness and Reidentifiability
14mBrowser Fingerprinting: Stalking With a Personal Touch