Deploying and Evaluating Pufferfish Privacy for Smart Meter Data

Realizing a Smart Grid without sacrificing the privacy of consumers is a challenging problem. Data-centric approaches like Puffer fish ensure privacy by transforming data so that certain user-specified information, so-called secrets, cannot be inferred. Deploying Puffer fish on smart-meter data requires application-specific decisions, i.e., A general definition of secrets in time series. We investigate how to perturb energy consumption data in this manner, and we quantify the trade off between privacy and utility.