Joe Konstan
University of Minnesota
Through the Experimentalist's Lens: Social Computing Research in Balance

"Big Data" is everywhere. And data from and about social computing systems keeps getting bigger. The wealth of available data leads to a natural temptation to seek knowledge in that data. This talk is both a warning and an illustration that big data isn't the be all and end all of research. Too often big data holds many answers, only some of which are true. And too often big data leads researchers to easy questions and away from hard but important ones. The GroupLens Research Group at the University of Minnesota has spent the past 20 years trying to balance data-centered research with experimental research. While perhaps best known for our MovieLens system (which has registered 200,000 users over the past 18 years), our group has carried out extensive system-building and online lab and field experiments across a variety of software system platforms, and through that work come to research conclusions impossible to derive from usage data alone. At the same time, we have produced datasets and software tools used by thousands of researchers and students, and that form the basis for nearly 1000 research papers by us and others. Drawing on examples from our past and current work in recommender systems, peer-contribution, crowdsourcing, and community knowledge building; I will lay out what we've learned about keeping big data analysis, laboratory studies, and field experiments and studies in balance.

Joseph A. Konstan is Distinguished McKnight University Professor, Distinguished University Teaching Professor and Associate Department Head of the Department of Computer Science and Engineering at the University of Minnesota. His research addresses a variety of human-computer interaction issues, including recommender systems, social computing, and applications of computing to public health. His work on the GroupLens Recommender System won the 2010 ACM Software Systems Award, and his publications include eight articles that have been cited more than 1000 times (Google Scholar Data). Professor Konstan has been recognized for his teaching through both University and College teaching awards. He has given popular webinars on recommender systems and on ethical issues in social computing research, and has taught dozens of short courses and tutorials on recommender systems, human-computer interaction, and related topics. He is currently teaching the second edition of his MOOC: Introduction to Recommender Systems. Dr. Konstan received his Ph.D. from the University of California, Berkeley in 1993. He is a Fellow of the ACM, IEEE, and AAAS; an elected member of the CHI Academy; Past-President of ACM SIGCHI, the 4500-member Special Interest Group on Human-Computer Interaction; a three-term member of the ACM Council; and co-Chair of ACM's Publications Board. He chaired he first ACM Conference on Recommender Systems in 2007, and also chaired ACM's UIST 2003 and CHI 2012 conferences. He is papers co-chair for the CSCW 2016 conference.