ampliation - Computer Science
Transcription
ampliation - Computer Science
Toward Interactive Visualization for Deeply Digital Scholarship Chris Weaver School of Computer Science University of Oklahoma Visual analytics research is increasingly focusing on deeply textured data and structured information. Analytics is poised to become the next general-purpose methodological science. Interactive visualization is the face of analytics. Visualization helps us share ideas about data. Interaction helps us express and explore them. Collaborating with scientists and scholars on realistic application of tool designs in a wide variety of knowledge domains is essential for methodological success. It doesn’t matter whether data is quantitative or qualitative, simple or complex, big or small. Is it interesting or useful for human understanding? There are many ways to visually encode data items, visually represent data collections, and interactively gesture to navigate and edit data from multiple perspectives simultaneously. Elucidating the possible and practical design space of interaction building blocks is the central theme of my visualization research. How can interactive visualization support a complete digital workspace for science and scholarship? Users Whether recorded directly by people or indirectly using machines, data captures our observations and interpretations of the world around us. Drag events and durations to shift them along a timeline. Drag things on a map to specify location, direction, or region. Developers Draw and erase arrows to add and remove value co-occurrences. Design Press modifier keys to select, negate, or otherwise classify values. Data Transformation Pipeline Interaction Sp ac e Data Edits Movement Highlighting ig av N ts ec Query bj at io n Side-Effects O We seek to expand the usefulness of interactive visualization tools by making them into meaningful graphical contexts for dynamically editing any data. Display s re tu ion es ct G Sele How can people interact with data to express their observations, interpretations, questions, and reasoning in a more direct and natural form? Pinch to associate uncertainty with a measurement or a set. Open-ended extension and modification of data in context. Populate new data sets by annotating values and associations. Explore idiosyncratic models of data relationships through tagging. Create, share, and assess competing hypotheses expressed as data. Views Interaction Mapping (interactions → data) Visual Mapping (data → graphics) raw interactions keyboard, mouse, wheel, multitouch, glove, wand, gaze... gestures keypress, click, drag, roll, swipe, pinch, fixation... data edits select, annotate, collect, compose + graphical side-effects cursor, rubberband, lasso, tooltip, menu... What We’re Working On ampliation annotation as a part of open-ended interaction physics chemistry a new language to question, observe, reason, conclude, share www.math.nyu.edu/~crorres/Archimedes/Lever/leverBigCorners.gif zoology Ampliative from Latin ampliare, "to enlarge" used in logic to mean “extending” or “adding to what is already known.” In Norman law, an “ampliation” was a postponement of a sentence to obtain further evidence. Our software workbench for creating visualizations is called Improvise. To see more, visit www.cs.ou.edu/~weaver/improvise urban planning psychology meteorology knowledge networks demographics archives economics political science history A comprehensive survey of the forms of visual data interaction in the research literature and in practice. • A new model of visualization interaction design that significantly extends the useful reach of visualization tools and facilitates their systematic and rapid development and implementation. • A new model of open-ended, column-based data organization that supports efficient transformation, representation, manipulation, and extension of data through interaction. • Understanding the data structures, transformations, representations, and manipulations needed to represent details of structured observations, reasoning, and interpretation. • Understanding how gestures can be effectively linked to visual encodings to support direct manipulation of data values of different types. • Learning how gestures can be composed to ask questions about data and make statements about observed patterns by extending and refining data. • Generalizable patterns and guidelines, grounded in evaluation, for designing visualizations that support exploration, analysis, and interpretation tasks through gesture composition. The research presented here has been supported in part by National Science Foundation Awards #1036331 and #1351055, and by a grant from the Andrew W. Mellon Foundation.