This research develops methods for interactive graphics, and makes them publicly available in the open source project, GGobi, freely available at www.ggobi.org. The development focuses on three areas: building the infrastructure; developing new methods for interactive facetting, tours and logical linking; and disseminating the research. Infrastructure development includes making the the data pipeline underlying GGobi more modular, improving the communication between GGobi and R, and adding more drawing primitives to facilitate map drawing. GGobi's capacity for handling categorical variables and large data is also expanded. The new methods development is driven by challenges provided in the analysis of multivariate longitudinal and spatio-temporal data. The project pays careful attention to disseminating the work, maintaining stability, portability and extensibility of GGobi, developing curriculum material and video demonstrations, and holding a yearly users and developers meeting.
This research extends an open source software project, called GGobi, which provides interactive plots to explore and discover hidden or unexpected structure in high-dimensional data. High-dimensional data arises in many, many real-life situations. For example, in environmental studies where weather balloons, or remote instruments mounted on satellites, measure variables such as temperature, humidity, wind speed, wind direction, and pressure at many locations on the globe, to study global warming. Each individual variable might not reveal much. Interactive graphics helps to piece together these multiple variables, to investigate different weather patterns, and discover the impact on local climate patterns. Other examples include studying the characteristics of images as might be used in homeland security applications, and exploring individual trends in longitudinal studies of health records or education achievement. The research pursued under this grant will extend the software to incorporate geographic maps, new methods for plotting categorical variables, and dealing with large amounts of data. As an open source software project a lot of energy is put into disseminating the work to the general community. The software is available freely so that it can be used by anyone with their own data, and that instructors can incorporate the software into their college-level classes. A web site with documentation and tutorials with demonstration movies is maintained at www.ggobi.org.
