Michelle Dowling is a DAC Ph.D. student in the department of computer science. She works at the InfoVis Lab, her advisor is Chris North, associate director of DAC. Her research focuses on how to allow users to either explicitly or implicitly gather knowledge about a set of data (contained in text-based documents, images, spreadsheets, or other various data sources) by interacting with a visualization. The visualization parameterizes each data source, and the sources are plotted onto a map using multi-dimensional scaling (MDS) algorithms. The parameters can then be upweighted or down weighted by the user to produce a different visualization.
Alternatively, the user may select points and move them further away from or closer to other points, which gives the algorithm feedback on which parameters should be more important, thus changing the visualization. This feedback or direct searches by the user gives the system information on how to query the database for more data sources to display on the screen. Other types of feedback include viewing the contents of the data source, highlighting important text, minimizing a data source, or deleting a data source from the visualization.
- Visual Analytics