Predict Behavior

Data extracted from social media, mobile apps, wearable sensors, and other technologies can be used to monitor and predict behaviors of individuals and groups. We are pioneering methods in this emerging field of “big data” research, by working with computer science, engineering, and simulation modeling groups to develop cutting edge methods to extract psychological information from “big data” sources to predict behavior.

For example, we collected over 150 million tweets on Twitter and identified tweets that suggested people might engage in HIV and drug-related risk behaviors. We modeled these tweets alongside a map of HIV cases in the United States and found that locations where people were tweeting about HIV and drug-use behaviors were more likely to have higher levels of HIV cases. This provides an approach where “big data” from technologies can be used for surveillance, detection, and prediction of behaviors and outcomes.

Pinmap of tweets