PITT Initiative on Computational Social Science
We are a research group at the University of Pittsburgh. Our mission is to understand human nature and society at scale.
Society generates, and is shaped by, more data than ever before. This creates the opportunity to study human nature and society in new ways. Nearly every field of social science stands to be transformed by data.
Learn more about our group
Request to join the mailing list, or to learn more about the PittCSS weekly research seminar: email mrfrank@pitt.edu
Research Themes
Learn about ongoing projectsOur group conducts research in Computational Social Science at the intersection of Data Science, Complex Systems, Economics, Sociology, and Political Science.
Economic Complexity
From international trade to workers' career to team management, economic systems rely on structure. Mapping and studying these structures builds on traditional economics by identifying the path-dependence of out-of-equilibrium dynamics. For example, relating international exports reveals the growth of national economies, occupations' skill requirements determine cities' economic resilience, and employee interactions determine information spread across a company, innovation, and the workers who are most central to corporate structure. Research in this area is enabled by large-scale data and methodologies following a broad spectrum of inspiration, including from physics, ecology, and complex systems.
Ethical & Accountable AI
Artificial Intelligence and machine learning (AI/ML) are increasingly used in public and private domains, ranging from business, public health, social services, to criminal justice, to assist in human decision-making. These AI technologies create tremendous ethical challenges in our society: biases are introduced in a number of ways through algorithmic decision-making aids, which can lead to biased outcomes, especially for minority populations and women. Addressing such challenges requires multi-disciplinary efforts. Research in this area will synthesize and innovate theories from machine learning, model building, ethic and cultural studies, psychology and cognitive sciences and lay the foundation to successfully navigate and evaluate responsible AI practices.
Science of Science
"Science of Science" aims to understand the properties, laws, and mechanisms of science creation. Over time, significant focus on evaluative metrics, especially ones based on citation, have become insufficient to capture the complexity of the evolution of science. With increasingly available data and computation, the recent Science of Science revival has provided a new array of whys, wherefores, and know-hows to expand the scope and depth of scientometrics by integrating theories and methods from sociology, economics, management science, psychology, and beyond.
Information Ecosystem
The ongoing pandemic marks the cusp of attention on how vulnerable people are being influenced by information distributed online. Problematic signals, from misinformation, disinformation, to polarized and biased messages, were seen to associated socially harmful attitudes and behaviors, such as discrimination and violence against disadvantaged, stigmatized or scapegoated groups. Research in this area attemps to understand all aspects of information ecosystem mediated by digital platforms, from powerful contents, vulnerable populations, to the chain of influence at scale.
News
- Kristina Lerman's Talk: Why Social Media is So Uniquely Toxic to Our Mental Health
- A Special Thank You to PittCSS Seminar External Speakers (Spring 2023)
- Nefriana and Alireza Presented at DINS PhD Student Speaker Series
- Yiling's article is accepted by Journal of Informetrics
- Guest Speaker Yong-Yeol 'YY' Ahn on Representation learning for computational imagination