Visualization for Managing Large Data Sets
December 4, 2011
Named for the quantitative visualization guru Stephen Few, the Thinking with our Eyes initiative aims to develop visual, dynamic analytic environments for large qualitative data sets. Student projects look to principles of visual encoding taken from the field of quantitative analytics and then apply them to qualitative data. This visual translation of data allows for faster, visual processing, allowing researchers to increase the size of the data set for analysis. Current student explorations use data from ethnographic inquiries conducted using online research platforms—internet-based technologies that use self-reporting techniques such as diaries and logging behaviors to collect cognitive and behavioral data.
This working prototype supports “data poking” or the rapid prototyping of qualitative data sets to highlight patterns at the macro level. Nineteen allows researchers to quickly display and interact with up to 1000 rows of data in .xlsx format. Nineteen uses color and position to cluster and track key variables. This aids analysis by visually compressing a multitude of data units into a single visual plane so researchers can see all their data at once, while allowing them to investigate each individual unit of data. Try it for yourself at data.pollari.org
Student Ted Pollari in collaboration with professor Kim Erwin on creating “small tools for big qualitative data.”
Visual Google applies visualization techniques to new kind of qualitative data set—search results. Student projects seek to transform linear, text-based search results into new, dynamic visual environments so as to increase the number of results accessible by end-users. This project explores how to give users access to content categories to allow them search and view results more directly using direct manipulation of category bubbles.
Student Helen Wills on “intertwingling” as a way to enable category-based search strategies.
Nineteen: Visualization, Prototyping, and Thinking With Our Eyes
Limelight: A New Way of Looking at Search