Computational Social Science
for Pandemic-Related Social Good
The COVID-19 pandemic has represented a massive shock to human society, driving profound and persistent changes across human cultural, social and economic activities. The digitalization of modern society has resulted in an explosion in digital traces and information, which (should) provide an evidentiary basis for consensus around data-driven policy to combat coronavirus and minimize its human costs. Nevertheless, the past three years have also manifested a pandemic of misinformation and disinformation, which have driven polarization in personal behavior and public policy. One cause underlying this conflict and confusion follows from the significant gap between scientific research and its public audiences. This suggests that one key to maximize the impact of scientific findings that promote social good, especially in the era of social media, is to make accurate findings and inferences relatable and transmissible across social networks.
In this datathon, you and your team of up to 5 participants are tasked with discovering and conveying insights from linked data about COVID prevalence, its framing in the national and local U.S. news, behavioral and commercial activities that follow. (A nonredistrubtion agreement will be required for proprietary data assembled exclusively for the datathon). We encourage teams to identify patterns and causal relations of scientific and societal importance and produce a clear, accurate, appealing and publicly relatable visualization and complementary textual description. In the datathon GitHub, we compose a data portfolio that covers the daily dynamic of Covid-19 cases, framings of the pandemic, patterns of human transportation, and interactions with business. Together, these depict a complex system that captures the co-evolution and feedback between social, economic and cultural domains. The datathon invites you to submit an original, publicly relatable impression of your discovered patterns or associations, which should include an artistically appealing and scientific accurate data-driven visualization of your findings, alongside an explanatory description of up to 500 words. Findings should relate at least two forms data described above, may but need not link additional information (e.g., Twitter activity, Google Search attention, etc.), and could explore a wide range of social and policy-relevant topics, including but not limited to the local and national propagation of misinformation about the pandemic; the impact of COVID framings on human social and commercial behavior; social and spatial predictors of COVID transmission; human costs stemming from the pandemic; and policy proposals simulated atop existing regional data.
Submissions will be evaluated by a committee of peers based on public relatability, societal importance, scientific novelty and artistic appeal. Participants will orally present their contributions in a “Datathon Slam” and we will also consider its propagation on real-world social networks (e.g., Twitter) over the course of the conference.
Prizes for the datathon include a $2000 prize for the best impression and description; a $1000 prize for the most innovative impression and description (conditional on accuracy); and a $1000 prize for the impression and description with the greatest public health implications. Winners will participate in a closing conference panel with Computational Social Science leaders. We will also give $25 Amazon gift card for all participants. Interested participants can sign up here below:
Fill out this form to signify participation.
We will then invite those signed up to a Slack playspace (firstname.lastname@example.org) where participants can find data playmates in the #findplaymates channel (if they don’t come with a prefigured team). Teams will create a separate private channel labeled by their team name (with all and only team-members invited) to claim and coordinate their datathon activities. Data will be released Monday, July 18 @ 8am CDT. Final submissions should be emailed to email@example.com by Wednesday @ 8am CDT. Rooms will available July 19 to coordinate and work on projects. (Some IC2S2 tutorials (July 19) are directly tuned directly to improve analyses with these data.)