2010-2011. MIT Media Lab.
Not everyone in online public settings seeks rich discussion. Often, casual users seek to quickly scan through a forest of interaction to get a quick understanding of what the contributors are saying; a difference that should be reflected in the design. But how can this be solved as the Internet increasingly inches towards Borges' Library of Babel (1941)? The high connectivity of the web affords an ever-increasing number of points of view, "true" facts, "false" facts, and tangential commentary and reactions for any given situation. Current media do not sufficiently allow easy comprehension of such a large amount of data, providing little context or summary. Reverse chronological and its approximations seem to be the staple of presentation. It is difficult enough to quickly ascertain the breadth of viewpoints that exist and their thought process or validity, let alone any community-centric information about the posters and what viewpoints are typical for them. Hiltz and Turoff (1978) long ago foresaw the desire and positive possibilities that could arise from shared online dialogue across heterogenous audiences. Now that this dream has become a reality in 2011, current asynchronous forums and commenting systems resort to paging long linear lists. With the increasing volume of opinions and article sharing, current interfaces aremust rethink the list.
Defuse attempts to do so, focusing on improving our ability to understand participants and the crowd. It uses various statistical and natural language processing methods to summarize subjects by their commenting history, and aggregates it even further for each article. Observers are given an interactive portrait of the crowd by the found demographics, that facilitates faceted drilling down into the raw comments. A data portrait of each author accompanies their messages. It demonstrates that machine learning of users’ digital footprints can facilitate the social and sociological navigation of crowds. These prototypes satisfy goals and curiosities that are political and demographic in nature.
Contained in Paper
Zinman, A. Me, Myself, and My Hyperego: Understanding People Through the Aggregation of Their Digital Footprints. PhD Thesis. Department of Media Arts and Sciences, Massachusetts Institute of Technology. September 2011.