A RAND-AUC Consortium Tech & Public Policy Hackathon
One way to think about the health of a society is in terms of the well-being of its most
vulnerable and disadvantaged members. Do they have access to the services they need?
To transportation? Are their neighborhoods safe? Do their jobs pay enough to support
them and their families? Are their rights protected? From the earliest days, it was clear
that the COVID-19 pandemic was making inequities worse, but its full impact on our
vulnerable communities is still coming into focus. The first step to effective policy analysis
is seeing a problem as clearly as the data will allow. The aim of this hackathon is to help
us see more clearly.
The guiding question for this hackathon is this:
Can we paint data-driven pictures of how our vulnerable populations fared
throughout the COVID-19 pandemic, and the extent to which government policies
(or lack thereof) have either increased or decreased inequities?
The goal is to apply your skills, ingenuity, and lived experiences to expose the inequitable
outcomes and effects of the pandemic, as well as the effects of our policy responses to
date. We encourage you to “center” the needs of vulnerable populations and incorporate
social justice considerations as much as possible in your explorations. To assist you in this
exercise, we will provide data sets that highlight different aspects of the U.S. response to
the pandemic, along with demographic data to help identify key sub-demographics.
Timeline: Oct. 27–Nov. 12, 2021 (16 days)
• Teams should set up regular team meetings/“stand-ups” once every two days
to brainstorm and make progress on their tasks.
• All teams will have access to a policy researcher mentor for consultation on analytic
ideas and useful resources. They will hold office hours during the week.
• Participants will be enrolled in dedicated Slack channels to enable in-team
communication (if needed, please refer to this Slack guide); Zoom will be used
for video conferencing for the interim review and the final presentation.
Teams will present their work at two events:
• Interim progress review
• Final presentation event
Initial data sets for analyses will be placed in a git repository. Teams are encouraged to
find other data sets as needed and create their own repositories (e.g., by cloning the git
repository) to contain their solutions and manage the code collaborations.
Scoping: Spend time with the data sets provided. Identify which vulnerable
populations and policy dimensions you want to explore. The types of analysis you
can undertake are constrained by the available data. You may also apply any external
data sources you can identify.
Suggested Approach: Pick a policy dimension of interest (e.g. health outcomes
education, voting). Discuss what you consider to be bad vs. good (or fair vs. unfair)
policy outcomes for the dimensions you select. Make sure you have data to inform
analyses on selected dimensions and/or sub-demographics.
Suggested Timing: Consider spending 2-4 days on this step for scoping and
Explore: Explore the problem using data and visualizations:
a. Explore effective ways of visualizing equity problems that involve
b. What do your visualizations and analysis hide? What might be missing?
Suggested Approach: Generate lots of tables and plots on the relevant data sets
that you have selected. Come up with ideas for relationships you are expecting
to see. Visualize and analyze the data to see if those ideas are supported.
Then Iterate! A lot!!! Once you have identified a data-supported story, flesh it out
with other data analyses and figures. You might need to find other data sources.
Talk to your mentors to help with these.
Suggested Timing: Consider spending 6-7 days to arrive at the main elements
of your first story that the data supports. You will have many false starts,
so consistency and repetition is the key. You have to look hard to see the
important patterns. Negative findings are also valid outcomes. You will likely
be refining your analyses until the end of the exercise.
Analyze: Demonstrate potential harms that may not be reflected in the
a. What kinds of policy problems could we miss if we only use recorded quantitative
data to inform our research?
b. What statistical methods, if any, can improve how well the data sets represents
the general population?
c. How would you go about trying to highlight perspectives missing from
Suggested Approach: Think in terms of “missing perspectives”! These can be
informed by your lived experiences over the course of the pandemic. What patterns
and outcomes did you observe in your pandemic experience?
Suggested Timing: Consider spending about 3 days on this question. But those
3 days should overlap with the exploration task as it can lead to deeper insights
for your analyses.
Propose interventions: Come up with some actionable steps or recommendations
to address the following questions:
a. What tools or approaches would be useful for addressing inequitable outcomes
in the area you have chosen to focus on? Make sure to look for past precedents
to help you qualify the novelty of your proposals.
Suggested Approach: Intervening effectively in complex policy systems while
avoiding unintended consequences is hard! Think carefully about what bad outcome
you are trying to prevent. Think also about what tools are available (regulation,
investment, communication strategies, etc.). If an intervention seems obviously
right, you are probably missing an important perspective or dimension…
Suggested Timing: About 3 days. This should be the tail-end of your work,
after you have really dug into the data and your findings.
Data Sets Available
Incidence Data Sets
|Signals||Data Source Options|
(incidence, mask use, prisons, colleges)
|GitHub COVID Repos: JHU CSSE, NYT|
|UCLA Prison Data||“COVID Behind Bars” Repo|
|Hospital Utilization + Capacities||HealthData gov portal|
Behavioral and Socio-Economic Data Sets
|Behavioral Signals||Data Source Options|
|Voting Behavior||Center for New Data
(in hackathon repo)
|PPP Loans||US SBA FOIA Site|
|Mobility Behavior||Mobility indices derived from
|COVID Tweets||Repo of Tweet IDs|
|COVID Vaccine Response||CDC Vaccine data set|
|2020 Census County-level Data||https://www.census.gov/
• Race + Ethnicity
• Median Income
• Health Insurance coverage
Teams are expected to document their process using a combination of
• narrative text (e.g., a short report)
• interactive code demonstrations (e.g., R markdown/RShiny, Jupyter notebooks)
• October 27, 5–6:30 pm ET – Kickoff
• November 5, 12–1:30 pm ET – Interim Review
• November 17, 4–6 pm ET – Final Presentations
TEAM 1 – AUC-cuity (watch presentation)
AUC Students: Abias Dotson (CAU), Melanie De La Rosa (Spelman), Paris Venable
(Morehouse), Zyahn Archibald (Morehouse), Ayden Clark-Veal (Morehouse)
Pardee RAND Students: Khrystyna Holynska, Carlos Villegas
Mentor: Jon Welburn
TEAM 2 – The Regulators (watch presentation)
AUC Students: Amari Torrance (Morehouse), Jemielle McMillan (CAU),
Michael Spurlock Davis (Morehouse)
Pardee RAND Students: Khadesia Howell, Keller Scholl, Catria Gadwah-Meaden
Mentors: Li Ang Zhang, Toyya Pujol-Mitchell
TEAM 3 – Can-You-Hack-It (watch presentation)
AUC Students: Jonathan Banks (Morehouse), Shane Brooks (Morehouse),
Elisha Azize (CAU), Lena Anglin (Spelman)
Pardee RAND Students: Eddie Lopez, Zet Wang, Priya Gandhi
Mentor: Gavin Hartnett
TEAM 4 – Healthy Hackers (watch presentation)
AUC Students: Cai Johnson (Morehouse), Jordon Mosby (Morehouse),
Mikyah Tabb (Spelman), Stephen Agyepong (Morehouse), Omarri Beck (Morehouse)
Pardee RAND Students: Krystyna Marcinek, Omair Khan
Mentor: Melanie Zaber
TEAM 5 – DataFree Thinkers (watch presentation)
AUC Students: Calvin Bell (Morehouse), Kendall Moore (CAU), Sade Fisher (Spelman),
Jalen White (Morehouse)
Pardee RAND Students: Swaptik Chowdhury, Nihar Chhatiawala
Mentor: Kathryn Edwards
TEAM 6 – Policy Pusherz (watch presentation)
AUC Students: Bennie Williams (Morehouse), Maya Griffin (Spelman),
Pamela Russell (Spelman), Amari Torrance (Morehouse)
Pardee RAND Students: Carlos Calvo Hernandez, Max Griswold, Zara Abdurahaman
Mentor: Jhacova Williams
• Angelica Geter, D.P.H., Chief Strategy Officer, Black Women’s Health Imperative
• Raynard Kington, M.D., Ph.D., Head of School, Phillips Academy, Andover;
Trustee, RAND Corporation; Board of Governors, Pardee RAND
• Michael Scholtens, M.S., Data Analyst, Digital Threats to Democracy,
The Carter Center
• Asya Spears, M.S., Ph.D. Candidate, Pardee RAND Graduate School;
Alumna, Spelman College
• Malcolm V. Williams, Ph.D., Senior Policy Researcher, RAND; Inclusion,
Diversity and Equity Advisor, Pardee RAND
Framing questions: Did the team develop exploration questions that were
interesting and provided viable scope for the hackathon?
Analysis: Did the teams use appropriate data sets and methods? Were questions
answered and/or did teams pivot when necessary? Did the teams visualize data
and/or results effectively?
Presentation: Did the teams clearly articulate (explain and provide support for)
what they did, how they did it, and why they made their analytic choices?
Did they discuss the problem, questions pursued, results of their work, and
Spread The Word
Share your hackathon experience on social media and let your networks know you’re
• Follow @RANDCorporation and @PardeeRAND
• Tag your posts on Instagram and Twitter with #HackingEquity
• Discuss what you’re learning and what you’re building
• Network with your team members and RAND mentors
• Crowdsource ways around your obstacles and share your successes
• Grab this image and share it 👇
If you have any questions, please reach out to either your mentor or
Hackathon Advisor and Judge Asya Spears (Pardee RAND Doctoral Candidate
and Spelman College alumna), [email protected].
For more information, see the Pardee RAND Hacking Equity page.