%0 Journal Article %J Big Data Research %D 2019 %T Interactive Visual Analytics for Sensemaking with Big Text %A Michelle Dowling %A Nathan Wycoff %A Brian Mayer %A Wenskovitch, John %A Leman, Scotland %A House, Leanna %A Nicholas Polys %A North, Chris %A Peter Hauck %K Big data %K interactive visual analytics %K Semantic interaction %K text analytics %K Topic modeling %K visualization %X Analysts face many steep challenges when performing sensemaking tasks on collections of textual information larger than can be reasonably analyzed without computational assistance. To scale up such sensemaking tasks, new methods are needed to interactively integrate human cognitive sensemaking activity with machine learning. Towards that goal, we offer a human-in-the-loop computational model that mirrors the human sensemaking process, and consists of foraging and synthesis sub-processes. We model the synthesis loop as an interactive spatial projection and the foraging loop as an interactive relevance ranking combined with topic modeling. We combine these two components of the sensemaking process using semantic interaction such that the human's spatial synthesis actions are transformed into automated foraging and synthesis of new relevant information. Ultimately, the model's ability to forage as a result of the analyst's synthesis activities makes interacting with big text data easier and more efficient, thereby facilitating analysts' sensemaking ability. We discuss the interaction design and theory behind our interactive sensemaking model. The model is embodied in a novel visual analytics prototype called Cosmos in which analysts synthesize structure within the larger corpus by directly interacting with a reduced-dimensionality space to express relationships on a subset of data. We then demonstrate how Cosmos supports sensemaking tasks with a realistic scenario that investigates the affect of natural disasters in Adelaide, Australia in September 2016 using a database of over 30,000 news articles. %B Big Data Research %V 16 %P 49 - 58 %8 July/2019 %U http://www.sciencedirect.com/science/article/pii/S2214579618302995 %R https://doi.org/10.1016/j.bdr.2019.04.003 %0 Generic %D 2018 %T Towards a Systematic Combination of Dimension Reduction and Clustering in Visual Analytics %A Wenskovitch, John %A Ian Crandell %A Ramakrishnan, Naren %A House, Leanna %A Leman, Scotland %A North, Chris %K Algorithm design and analysis %K clustering %K Clustering algorithms %K Data visualization %K Dimension reduction;algorithms %K Manifolds %K Partitioning algorithms %K Visual Analytics %K visualization %B IEEE Transactions on Visualization and Computer Graphics %V 24 %P 131-141 %8 01/2018 %R 10.1109/TVCG.2017.2745258 %0 Conference Paper %B Proceedings of the International Working Conference on Advanced Visual Interfaces %D 2012 %T The semantics of clustering: analysis of user-generated spatializations of text documents %A Endert, Alex %A Fox, Seth %A Maiti, Dipayan %A Leman, Scotland %A North, Chris %K clustering %K text analytics %K Visual Analytics %K visualization %X Analyzing complex textual datasets consists of identifying connections and relationships within the data based on users' intuition and domain expertise. In a spatial workspace, users can do so implicitly by spatially arranging documents into clusters to convey similarity or relationships. Algorithms exist that spatialize and cluster such information mathematically based on similarity metrics. However, analysts often find inconsistencies in these generated clusters based on their expertise. Therefore, to support sensemaking, layouts must be co-created by the user and the model. In this paper, we present the results of a study observing individual users performing a sensemaking task in a spatial workspace. We examine the users' interactions during their analytic process, and also the clusters the users manually created. We found that specific interactions can act as valuable indicators of important structure within a dataset. Further, we analyze and characterize the structure of the user-generated clusters to identify useful metrics to guide future algorithms. Through a deeper understanding of how users spatially cluster information, we can inform the design of interactive algorithms to generate more meaningful spatializations for text analysis tasks, to better respond to user interactions during the analytics process, and ultimately to allow analysts to more rapidly gain insight. %B Proceedings of the International Working Conference on Advanced Visual Interfaces %S AVI '12 %I ACM %C New York, NY, USA %P 555–562 %@ 978-1-4503-1287-5 %U http://doi.acm.org/10.1145/2254556.2254660 %R 10.1145/2254556.2254660