Skip to content
Ken Haase edited this page Sep 27, 2024 · 3 revisions

Welcome to the ddhca-release wiki!

This repository contains source code for the SafeTab-P disclosure avoidance application. SafeTab-P was used by the Census Bureau for the protection of individual 2020 Census responses in the tabulation and publication of the Detailed Demographic and Housing Characteristics File A (DDHC-A).

Using the mathematical principles of formal privacy, SafeTab-P infuses noise into Census survey results to create privacy-protected microdata which are used by Bureau subject matter experts to create the 2020 DDHC-A product. SafeTab-P was built on Tumult's "Analytics" platform and both SafeTab-P and the underlying platform are implemented in Python. The latest version of the platform can be found at https://tmlt.dev/.

In the interests of both transparency and scientific advancement, the Census Bureau committed to releasing any source code used in creation of products protected by formal privacy guarantees. In the case of the the Detailed Demographic & Housing Characteristics publications, this includes code developed under contract by Tumult Software (https://tmlt.io/) and MITRE corporation. Tumult's underlying platform is evolving and the code in the repository is a snapshot of the code used for the DDHC-A.

Warning: pyarrow library vulnerability

There is a known vulnerability in the binary Python wheel at: tumult/core/tmlt_core-0.6.0-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

This wheel contains a binary version of the pyarrow module with a vulnerability identified in 2023 (https://nvd.nist.gov/vuln/detail/CVE-2023-47248). The SAFETAB-P code itself does not expose this vulnerability, but any modified or extended versions should rebuild the wheel from sources beneath the tumult/core prefix or later sources from Tumult itself.

Clone this wiki locally