Investing Data with Untrusted Parties using HE

Data such as an individual's income, favorite sports team, typical commute route, vehicle maintenance history, medical records, etc. are typically not useful for making large-scale decisions such as where to build a new hospital, identifying which roads are in need of upkeep, and the like. However, aggregates of of these data across hundreds of individuals are useful to governments and to companies. Data cooperatives/unions offer a place for individuals to store their data and a service of data aggregation and interpretation to governments, non-profit organizations, and businesses while maintaining individuals' anonymity. We propose the use of anonymization techniques coupled with graph algorithms over homomorphically encrypted (HE) graphs as a basis of analysis for this accumulated data. We believe this approach ensures individuals' privacy and anonymity while preserving the usefulness of the plaintext data.