As information consumers, the internet has radically changed how we individually live our lives. However, we are also information producers, broadcasting and storing a great deal of information about ourselves on the internet. The fundamental question I am interested in is how we can use these troves of online information to measure and model large-scale human behavior.
In this initiative, we’re investigating ways to use online data to learn the distribution of various traits (e.g., gender, age, education) in online populations. Our focus is currently on using the Twitter platform, though we are now extending our work to others as well.
One of the long-term goals for this work is to develop techniques for mapping the demographics of online populations onto their real-world counterparts. We have some exciting, on-going work that shows that it is possible, for example, to infer the gender makeup of certain urban communities using only Twitter data.
Situational and Emotional Awareness
In times of crisis or turmoil, it is important for responders and effected populations to have accurate and current information to act on. Through funding from the Canadian government, we are investigating techniques for deriving both situational and emotional awareness from social media.
Controllability of Complex Systems
In an ongoing collaboration with a control theorist, we are investigating how complex systems, represented as networks, can be controlled – influenced to behave in one way or another. This is an exciting direction at the intersection of control theory and network science.