Seminar: Pipelines for Computational Social Science Experiments and Model Building: Exemplars of Mechanistic and Data-Driven Models to Explain Human Behavior
Vanessa Cedeno-Mieles
Postdoctoral Research Associate
Environmental Institute, University of Virginia
Friday, April 11
9:00 - 10:00AM
1100 Torgersen Hall
Abstract
Networked social science experiments are gaining significant attention as a means to understand human behavior at scale. Conducting these experiments and analyzing their data requires substantial effort, especially in computational modeling and automated analysis.
In this talk, I will present the design and development of composable, extensible software pipelines that automate the evaluation of social phenomena through iterative experiments and modeling. While this combined approach has been applied in studies, it is often done manually or conceptually. I will discuss how these pipelines were used to explore intra-group cooperation within a controlled, web-based networked game. To predict player behavior, I developed mechanistic and data-driven models of human decision-making. The agreement between model predictions and experimental data not only helps scientists explain observed behaviors but also opens new avenues for generating insights into complex social dynamics.
Biography
Vanessa Cedeño-Mieles is a Postdoctoral Research Associate at the Environmental Institute at the University of Virginia. Her research focuses on machine learning, data mining, simulations and networked online social science experiments. Through controlled online networked temporal social science experiments she studies social behavior, uncovering valuable insights into social phenomena through data analysis, simulation and modeling. With over 10 years of academic experience in Latin America, Vanessa is passionate about creating opportunities in Computer Science for students and communities by mentoring and fostering a culture of academic excellence and innovation. She earned her Ph.D. in Computer Science from Virginia Tech.