Adapting Differential Privacy in Scenarios

Privacy leakage events have been increasingly frequent recently. Your DOB, home address, workplace, household composition, race, shopping and movie preferences, and other sensitive information can be easily compromised without your awareness. Differential Privacy is a provable robustness approach to address those privacy issues. This thesis proposed three Differential Privacy methods to protect location data and census data. First method can protect your accurate home address while maintaining its South East position to the CBD. Second method can protect COVID infected patients’ home addresses and meanwhile preserve the distribution of those cases for epidemiological study. Third method protect census data.