Picture the Future: Graphical
Innovation in Environmental Statistics
Peter Guttorp
Many problems in environmental statistics
deal with space-time data. There is a serious discrepancy between the tools
we have for exploring data on the computer, such as point cloud rotation,
brushing and other linked plots, virtual reality displays, and other dynamic
computer displays, and the very limited static displays, at best allowing
color, that are allowed in journals and books. More and more are we faced
with large data sets, and the development of tools for handling them are
of high priority in the statistical computing community.
I will discuss some aspects of current
research in statistical graphics as it pertains to environmental data.
Specifically, I will talk about:
The talk will be illustrated with examples from current work at the National Research Center for Statistics and the Environment.
- Display of uncertainty in contour lines
- MCMC estimation of heterogeneous spatial covariance
- Looking for structure in spatially located multivariate data
- Dynamic displays of space-time processes
PETER GUTTORP
National Research Center for Statistics
and the Environment
Box 351720
University of Washington
Seattle, WA 98195-1720, USA
peter@stat.washington.edu
Identification of Outliers
and Homogeneous Groups in Very Large Data Sets
Ali S. Hadi
"Whoever knows the ways of Nature
will more easily notice her deviations; and, on the other hand, whoever
knows her deviations will more accurately describe her ways."
Francis
Bacon, Novum Organum II 29, (1620).
Data, which come from many fields, often contain outliers. Most statistics methods assume homogeneous data in which all data points satisfy the same model. However, as the aphorism above illustrates, scientists and philosophers have recognized for at least 380 years that real data are not homogeneous and that the identification of outliers is an important step in the progress of scientific understanding because outliers can distort the results of the analysis, and hence the conclusions based on these results. In this talk, I shall present recently developed methods for the identification of outliers in very large data sets.
ALI S. HADI
Department of Statistical Sciences
Cornell University
358 Ives Hall
Ithaca, NY 14853-3901, USA
ali-hadi@cornell.edu