Mattias Rost, Louise Barkhuus, Henriette Cramer (Yahoo), Barry Brown, “Representation and communication: Challenges in interpreting large social media datasets”, will be presented at CSCW’13 in San Antonio, Texas, in February. CSCW is one of the most highly ranked conferences in human-computer interaction and related fields. Acceptance rate is very competitive and only 5% of the papers are nominated for Best Paper. This work was performed at the Mobile Life Centre, in Stockholm, Sweden, just before Henriette joined Yahoo Labs. Mobile Life’s Barry Brown is a collaborator with Yahoo Labs through the Faculty Research Engagement Program. First author Mattias Rost will be presenting in San Antonio. The paper reflects on using social media data in understanding human behavior. It highlights the importance of understanding how people use social media to communicate with each other, as this affects the large-scale data that they generate and the results of the analysis of such data. This is especially important now, when the trend is to focus on big data, and on collecting more and more user-generated data to do large-scale analyses. Web-scale data analysis offers a wealth opportunity, but we need to remember We need to not forget the people who actually generated all of this data points, and whose behaviors are reflected in the data – and who, and what are missing from it!. We need to deeply understand their motivations, their social context, the decisions they make, and what they do. The thoughtful combination of the big and the small, qualitative in-depth studies of individual users and their habits and motivations is what is going to make web-scale studies of user-generated data a success. This means that we should not only focus on the BIG, but also appreciate the role of qualitative, in–depth studies and observations on a human scale. Bio: Henriette Cramer is a Research Scientist at Yahoo Labs. Her research focuses on mobile and location-based interactions, people’s perceptions of their local surroundings, ’Research in the Large’ methodological issues in using wide distribution of apps and web-scale data in research, and users’ interaction with autonomous and adaptive ’things.’