Census wants help evaluating its privacy-preserving software
The Census Bureau is looking for help auditing software that preserves the privacy of data the agency will collect and analyze in the 2020 population count.
In a request for information, the Census Bureau said the open source software it's developing for the count is designed to work with confidential data and will perform privacy-preserving data analysis. But because it hopes to freely redistribute its software with test datasets, it wants a third party to audit the software for privacy vulnerabilities. The software will be used for the 2020 Census of Population and Households, the 2017 Economic Census and other data products.
The vulnerabilities Census wants to find include the improper mingling of sensitive and non-sensitive data, the movement of data from the sensitive category to the non-sensitive category without a privacy analysis and the improper implementation of a privacy-preserving algorithm, according to the RFI.
The agency said its software will be written in the Python programming language, maintained in a git-based software repository, run on a Linux server an may possibly use the Apache Spark data analysis platform to achieve scalability.
More information is available here.
Connect with the GCN staff on Twitter @GCNtech.