U.S. Broadband Coverage Data Set: A Differentially Private Data Release

TL;DR

This paper presents a publicly available U.S. broadband coverage dataset at the zip code level, utilizing differential privacy to protect household privacy while providing error estimates for data accuracy.

Contribution

It introduces a novel broadband coverage dataset that applies differential privacy and includes error range estimates to ensure privacy without compromising data utility.

Findings

Differential privacy effectively preserves household privacy.

Error range estimates quantify data accuracy.

Dataset supports research on broadband access disparities.

Abstract

Broadband connectivity is a key metric in today's economy. In an era of rapid expansion of the digital economy, it directly impacts GDP. Furthermore, with the COVID-19 guidelines of social distancing, internet connectivity became necessary to everyday activities such as work, learning, and staying in touch with family and friends. This paper introduces a publicly available U.S. Broadband Coverage data set that reports broadband coverage percentages at a zip code-level. We also explain how we used differential privacy to guarantee that the privacy of individual households is preserved. Our data set also contains error ranges estimates, providing information on the expected error introduced by differential privacy per zip code. We describe our error range calculation method and show that this additional data metric does not induce any privacy losses.