UChicago Hosts International Computational Social Science Conference, Emphasizing Collaboration and Public Impact
In the last decade, new techniques for gathering, analyzing, and extracting insight from large-scale data have revolutionized social science, enabling new ways of examining our world. Working in tandem with computational experts, economists, sociologists, political scientists, psychologists, and other investigators of social dynamics can now harness a flood of new data from digital sources into large-scale, distributed experiments that were previously impossible.
Those innovative collaborations were in the spotlight as more than 500 scholars from around the world came to Chicago in July for the 8th International Conference on Computational Social Science (IC2S2), the young field’s premier gathering. The event, held at the Harper Center from July 19-22, cut across disciplinary boundaries with participants from around the world to engage both computer and social scientists in conversations about research and the technical approaches that can unlock future discoveries.
“IC2S2 was formed as a multidisciplinary conference to provide a sustained conversation between social and computational scientists,” said James Evans, Max Palevsky Professor of Sociology and Founding Faculty Director of the Computational Social Science program at the University of Chicago. “The magic of it is social scientists and computer scientists working together to address major societal issues ranging from discrimination in the workplace to diverse, resilient communities to how we support societies to equitably support health and share vaccines.”
The Chicago conference was co-organized by Knowledge Lab, the Center for Applied Artificial Intelligence, and the Mansueto Institute for Urban Innovation. The four-day event provided a mix of programming, with tutorials in computational approaches including deep learning, computer vision, and tools for running massive, online experiments, research talks and panels, and a “datathon” where teams formed to derive actionable, policy-relevant insight from large-scale data on COVID morbidity and mortality, local and national COVID news, human mobility and commercial engagement during the pandemic.
Digital Tools and Insights for Online and Offline Worlds
The conference’s keynote speakers included experts from government and academia on topics ranging from literature, censorship, urban infrastructure, and political polarization to artificial intelligence, natural language processing, and cognitive psychology.
Opening speaker Audrey Tang, Taiwan’s Digital Minister, described social technologies that empower diverse collaboration between the country’s government and its citizens. For example, they use a platform called pol.is that allows people to comment on and directly influence policy issues such as the introduction of ride-sharing services to Taiwan. Tang also focused on efforts to protect digital human rights, protecting personal privacy while still capitalizing upon the democratic promise of online social platforms. One idea: “soul-based tokens,” inspired by the video game World of Warcraft, that could enable community participation and voting rights without manipulation by bots or trolls.
On the other side of the battle over digital freedom, Molly Roberts discussed tensions in China over the Internet, where business and political interests clash and censorship is extensive. Roberts used social media data and surveys to study the sturdiness of the “Chinese Firewall,” which blocks access to sites based outside the country. Her research found that more people attempt to circumvent this censorship during times of crisis such as the early months of COVID-19, but that it largely functions to “tax” attention and reduce awareness on topics, such as politics, in which most have tepid interest.
Chris Bail, Professor of Sociology and Public Policy at Duke University, talked about using social media data to study polarization, including the design of a bespoke research platform, called DiscussIt, to discover how anonymity and gender influence political disagreement. Rediet Abebe, Assistant Professor of Computer Science at the University of California, Berkeley, described the myriad issues with software-based technologies such as DNA sequencing and “Shotspotter” monitoring when they are used in the justice system, and described a new auditing framework for these tools that can be employed by defense lawyers to determine if they were used appropriately.
Others focused on powerful new methodologies available to social scientists in today’s data-rich environment. Tom Griffiths, Henry R. Luce Professor of Information Technology, Consciousness and Culture, and director of the Computational Cognitive Science Lab at Princeton University, described his “manifesto” for a computational cognitive revolution, moving from asking binary questions to understanding more complex relationships across varied contexts. Whether it’s re-confirming and deepening classic psychological concepts such as the Dunning-Kruger Effect or designing new types of experiments, computational methods mean that “Rather than just asking does X affect Y, you can ask how does X affect Y,” Griffiths said.
“I definitely got more ideas per hour than I have at any recent conference,” Evans said. “Unlike a traditional conference, it’s really about bringing people that are far apart but within this shared space of digital social data, thinking about how we can understand and improve society and culture. It’s one of those conferences where it’s designed to really stretch the way in which people are doing research, pushing the work to be ambitious in new ways.”
Datathon Excavates Actionable Discoveries From Pandemic Data
As IC2S2 attendees heard about recent advances in computational social science, many also took part in a live research project run concurrently as part of the event’s Datathon. Titled “Computational Social Science for Pandemic-Related Social Good,” this activity challenged volunteer teams to discover new insights and patterns about the social effects of COVID-19 using data provided to the conference by the New York Times, Meta, Proquest, TDM Studio, and SafeGraph.
Forty teams entered the competition, and the finalists were invited to present their findings on the main stage of the conference before attendees voted for three awards: best overall, most creative, and the research with the greatest public health impact. The best overall project was conducted by an international team of scientists from Singapore, Japan, and the U.S., who used statistical methods to examine the factors, such as age and acceptance of mask-wearing, behind the wide variance in COVID fatality rates in different cities.
The most creative prize went to an undergraduate group from the University of Chicago that studied how the sentiment of COVID news affected foot traffic to businesses in Chinatowns across the U.S. relative to other ethnic neighborhoods in Chicago and other cities as a measure of anti-Asian sentiment during the pandemic. The team that received the public health prize was commended for using sentiment analysis and partisan polarization data to explore differences in COVID responses between “red” and “blue” states.
In keeping with the conference’s focus on public impact, the teams were judged not only on the sophistication of their methods and analysis, but also on how the results were presented.
“When you are doing social science, you should also think about how this really can have a social impact,” said Fengli Xu, a postdoctoral fellow at the Mansueto Institute for Urban Innovation who directed the Datathon. “We asked researchers to convey an important message in a publicly relatable way, to emphasize the social aspect of computational social science research.”
IC2S2 2022 was supported by Chicago Booth, the Data Science Institute, the Social Sciences Division, the Office of Research and National Laboratories, the Institute for Mathematical and Statistical Innovation, the National Science Foundation, the Alfred P. Sloan Foundation, Meta, ProQuest, and Sage Publications. That broad representation from several different UChicago initiatives touching upon computational social science underscored the importance of hosting such a prestigious and cutting-edge conference on campus, Evans said.
“I think there’s a ton of foci around computational social science, from the Data Science Institute’s Data & Democracy initiative and Booth’s Center for Applied Applied AI to the range of programs now mixing social and computational science in The College, from the data science major to our new computational social science minor,” Evans said. “The conference was a natural place to showcase those initiatives and the University’s commitment to this transformative and growing field.”