Aplicação de L-Diversidade na anonimização de dados públicos da campanha de vacinação contra COVID-19.

Abstract

Dissemination of data is a process that occurs with the goal of bringing more transparency and making data analyses enabled in general. Aiming to ensure the privacy of data, many disclosures are made by anonymizing (fat data bank) recordings from removing information that identifies the individuals involved, as is the case for disclosures of public vaccination data against COVID-19. However, there are attacks that can be easily performed on anonymized data only by associating registrations, through common attributes with other data disclosures with identifiers that do not have sensitive information. For this, several anonymization techniques have been developed as, for example, L-Diversity. This paper aims to show the privacy gain by applying that technique on vaccination data, in which association attacks were conducted using the profile of PROUNI beneficiaries and public scheduling data disclosed by the Fortaleza municipal prefecture. As a result, it was possible to observe a substantial increase in protecting sensitive information.