Wesley Willett, University of California, Berkeley
Jeffrey Heer, Stanford University
Maneesh Agrawala, University of California, Berkeley

Social data analysis – in which large numbers of collaborators work together to make sense of data – is increasingly essential in a data-driven world. However, existing social data analysis tools don’t consistently produce good discussion or useful analysis. Our recent work shows how analysts can enlist paid crowd workers to systematically generate explanations for trends and outliers at scale and demonstrates techniques for encouraging high-quality contributions.

In our workflow, analysts automatically or manually select charts showing trends and outliers in a dataset, then use crowd workers to iteratively generate and rate the best candidate explanations for those features.

Continue reading →

Travis Kriplean, Computer Science & Engineering, U. Washington
Michael Toomim, Computer Science & Engineering, U. Washington
Jonathan Morgan, Human Centered Design & Engineering, U. Washington
Alan Borning, Computer Science & Engineering, U. Washington
Lance Bennett, Political Science, Communication, U. Washington
Andrew Ko, The Information School, U. Washington

Communication is about listening as much as speaking. Unfortunately, our web interfaces have thus far paid scant attention to supporting listening, creating a feedback gap and likely contributing to the scorched earth nature of our web dialogue. We have designed Reflect, a simple interface for encouraging listening, and deployed it on Slashdot. As a by product of their acts of listening, commenters open up new possibilities for crowd-sourced discussion summarization.

Continue reading →

Twitter users choose to follow “insiders” of fields they are interested in knowing, even though they do not necessarily know them in person. External recommendation system that suggests expert accounts of a particular field for users to follow is a useful tool for this purpose. However, we questioned the accuracy and overall user experience of such system by analyzing the topical relevance based on users’ published contents (aka, Tweets), which is the major trend of Twitter recommendation system in use.

Continue reading →

Svetlana Yarosh, Georgia Institute of Technology
Tara Matthews, IBM Research — Almaden
Michelle Zhou, IBM Research — Almaden

Expert finding systems help enterprise workers find a person who may be able to help with a project or question, but it is challenging to select the best expert for a specific task. This leads the user to waste time by asking an expert who may not be able to answer the question or who might not take the time to respond (we call this a “strikeout”). Through a lab study of realistic expert-finding tasks with 35 participants, we found that the current leading expert finding approach leads to a 28% strikeout rate, but our improved interface reduces this rate to 9%.

Continue reading →

Anthony Tang, University of Calgary
Sebastian Boring, University of Calgary

While watching live sports events, many fans use social media (e.g., Twitter) to report on their thoughts about the event as it unfolds. We investigated using this information source to semantically annotate live broadcast sports games (here: American Football) to select video highlights that are appropriate for an individual fan rather than for a general audience – showing plays that are (1) interesting, (2) controversial, or (3) skillful.

Continue reading →

Chang Hu, University of Maryland
Philip Resnik, University of Maryland
Yakov Kronrod, University of Maryland
Ben Bederson, University of Maryland

In our previous post, we introduced MonoTrans, a system that does crowdsourced translation without bilingual people. In our research papers we presented some promising results using it to translate children’s books and Haitian Earthquake SMS messages. An interesting problem is: beyond experiments, how do systems like this work in the real world?

Continue reading →