The Federal Communications Commission (FCC), advocacy groups, scholars, and others interested in Internet performance have had wide-ranging discussions over the years on two related (but distinct) topics: (1) benchmarking Internet performance standards (“Am I getting what I’m paying for?”); and (2) equity (“digital divide” challenges, Internet access haves and have-nots). Various organizations, from regulators to public interest groups, have developed tools, methods, and datasets to try to answer these questions.
In 2021, the Infrastructure Investment and Jobs Act authorized $65 billion to expand broadband access and adoption, and address disparities highlighted by the COVID pandemic and the shift to remote work, school and health care. But to invest that funding most effectively for improving digital equity in the United States, new and deep data at a local scale must be gathered and analyzed to drive effective policy and advocacy solutions for the citizens and neighborhoods most in need.
In this exercise you will explore how Internet performance in a specific high-income neighborhood compares with Internet performance in a specific low-income neighborhood.
We have been performing network performance measurements across nearly 30 neighborhoods across Chicago over the past year to try to get a better handle on some aspects of how performance varies across neighborhood and time. Load some of the into a program you are comfortable performing some analysis in. We have pre-loaded some of the data into the following Google sheets as read-only. You may need to make a copy to perform some analysis.
Take one of the files of interest and see if you can find any trends. For example, how does throughput vary over time? How does latency or latency under load correlate with throughput?
(The above links are accessible from the University of Chicago. Public data is also available from the project page.
The devices_by_community.csv file has some information pertaining to devices
by community. One could perform a neighborhood-by-neighborhood comparison of
these devices to see how performance varies across communities, for different
dimensions.
We have performed some versions of that analysis in this paper where you can read more about these comparisons. The data we have provided also allows you to perform some of that analysis yourself.
Some simple computations you could explore include:
- What is the average throughput (or latency, or DNS lookup time) for a device in Logan Square? In South Shore?
- What is the device with the lowest average throughput in each neighborhood? What is that value?
Approaches such as those you have explored in this assignment could provide researchers and policymakers an opportunity to offer citizens a data-driven approach to address their core concerns about Internet equity.
How would you frame and formulate questions concerning equity and “digital redlining” in ways that could be suppoted by data analysis?