Location

MIT Media Lab, E14-240

Description

In this talk, Peter Sheridan Dodds and Chris Danforth will discuss their work on building what they call "lexical meters"—online instruments that use social media and other texts to quantify population rates of a wide array of human behavior such as wealth, exercise levels, obesity rates, and sleep insufficiency. They will first showcase their hedonometer, an instrument for measuring positivity in written expression, showing how they have consistently improved their methods to allow them to explore collective, dynamical patterns of happiness found in massive text corpora including Twitter, song lyrics, works of literature, movies, political speeches, and news sources. They will present evidence for how 10 diverse natural languages appear to contain a striking frequency-independent positive bias, describing how this phenomenon plays a key role in their instrument's performance, and how it may more deeply reflects human nature. They will then discuss their work on building the Panometer, introducing their latest instrument: the Lexicocalorimeter, a principled meter that turns phrases into calories. Finally, they will point to a number other diverse projects being carried by their team in the Computational Story Lab, ranging from the stories of sports to the dynamics of climate change.

Biographies

Peter Sheridan Dodds and Chris Danforth co-direct UVM's Computational Story Lab, a group of applied mathematicians and data scientists at the undergraduate, masters, PhD, and postdoctoral level working on large-scale, systems problems in many fields including sociology, nonlinear dynamics, networks, ecology, and physics. The group’s methods encompass sociotechnical experiments, data collection and analysis, and the formulation, analysis, and simulation of theoretical models. Dodds is the Director of UVM's Complex Systems Center. Dodds's training is in theoretical physics, mathematics, and electrical engineering with extensive formal postdoctoral and research experience in the social sciences. Danforth is the Flint Professor of Mathematical, Natural, and Technical Sciences. His formal background is in nonlinear dynamics applied to weather and climate prediction, and he is a member of the Mathematics & Climate Research Network. They maintain general research and teaching interests in complex systems and networks with a current focus on sociotechnical and psychological phenomena including collective emotional states, contagion, language, and stories. Dodds & Danforth have received funding from NSF, NASA, ONR, and the MITRE Corporation, among others, with Dodds being awarded an NSF CAREER by the Social and Economic Sciences Directorate.

Host/Chair: Iyad Rahwan