TNL Hackathon Explores Ethics of COVID-19 Dashboards

September 12, 2020

Data dashboards focused on COVID-19 are seeing increased use both by the public and by decisionmakers. Development teams and data scientists tend to focus on issues of data quantity and quality, but they rarely address ethical questions and elements of bias.

This summer, Pardee RAND's Tech + Narrative Lab (TNL) — with the support of the COVID Alliance, a nonprofit organization building a research collaboration platform for COVID-19 response, and the Public Interest Technology University Network (PIT-UN) — built on its previous policy hackathon experience to explore the ethics of data dashboards.

Hackathons are traditionally cybersecurity or machine learning competitions. The TNL’s policy hackathons are structured as team-based competitions involving activities such as proposing novel policy solutions to a vexing social dilemma, or analyzing a dataset with policy relevance for insights, and to create helpful visualizations.

A core tenet of the TNL is to consider both the applications and implications of new technology, and the ethics hackathon offered a structured format for students to consider these issues while showcasing how ethics can be integrated throughout the Pardee RAND curriculum.

“This hackathon was unique in the ethics framing combined with the big data aspects,” said TNL director Todd Richmond. “The groups were able to ask different questions, and identify gaps and what data wasn’t, but should be, collected.”

Students were asked to consider the ethical implications — such as potential harms to vulnerable groups — of the COVID Alliance's new research collaboration platform, which includes live geolocation data, hospital capacity data, infection data, and other information that U.S. decisionmakers need to enable public health interventions. Several Pardee RAND students are members of the COVID Alliance and initiated the collaboration.

The hackathon offered a mutually beneficial partnership. The COVID Alliance received feedback on the platform related to ethical considerations they may have not considered. In turn, hackathon participants gained practical experience advising a receptive partner while demonstrating their broad analytical skillsets.

Three teams — comprising 12 students, one alum, and four faculty advisors — worked over four weeks this summer.

The students who participated were Hamad Al-Ibrahim (Cohort '18), Pau Alonso Garcia-Bode ('18), Jalal Awan ('17), Sangita Baxi ('17), Jarrett Catlin ('18), Joan Chang ('18), Eddie Lopez ('20), Karishma Patel ('17), Nabeel Qureshi ('18), Keller Scholl ('19), Max Steiner ('19), and Sara Turner ('15). Recent alum Meg Chamberlin ('14) also participated, as did faculty members Osonde Osoba, Ben Boudreaux, Tepring Piquado, and Tricia Stapleton.

Each team sought to answer the question, “What are the potential ethical harms of the COVID platform and how can they be mitigated through design features, visualizations, and use-restrictions?”

The hackathon included an initial kickoff meeting, weekly office hours with the COVID Alliance, and a final meeting where teams showed the results of their efforts.

Teams performed a variety of data visualizations, particularly focused on missing populations and privacy concerns. They also explored how public health decision based on data with missing components might have unexpected adverse effects.

—Monica Hertzman