In the past two years we have observed an explosion in the use of online message streams. With an abundance of information comes information overload and a scarcity of attention: Active message stream users now face thousands of unread messages every day in their personal streams. To date, we are aware of little academic research dedicated to this stream consumption problem, despite the enormous interest among users and businesses.
This research will establish scientific foundations for our understanding of the consumption of message streams, and engineer, deploy, and evaluate new tools for users built on those foundations. This research has two principle aims. First, construct computational models that characterize the individual preferences of stream readers and predict for each reader the value provided by specific messages, and, in aggregate, specific stream contributors. The second aim is to design novel, personalized and intelligent user interfaces that facilitate better digestion and management of online message streams, and evaluate those designs in large-scale field studies on the popular platforms Twitter, Facebook and Google Reader. The proposed research is among the first to examine online message streams broadly and to characterize message streams from the perspective of consumption. Our modeling will give insights into the nature of online message stream consumption and provide a theoretical framework for further investigation.
Hundreds of millions of users worldwide are ?voting with their fingers? to spend time creating, commenting on, communing over, and consuming tweets, wall posts, status updates, and blog posts. The research proposed here promises to deepen our understanding of the purpose, value, and social mechanisms underlying message streams, and to create new computational resources for modeling the contributors, their contributions, and the process of consumption. Based on these computational models, innovative visualizations, recommendations, and interfaces will be developed to reshape the user experience, and inform the design of future message stream services.