|
Activity Number:
|
157
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Graphics
|
| Abstract - #307590 |
|
Title:
|
Data Display Principles Revealed in the NASA Data
|
|
Author(s):
|
Rafe Donahue*+
|
|
Companies:
|
Vanderbilt University Medical Center
|
|
Address:
|
1716 Danforth Park Close, Brentwood, TN, 37027,
|
|
Keywords:
|
data display principles ; data visualization
|
|
Abstract:
|
The complex nature of the NASA weather data certainly makes it difficult to understand. The data are a multivariate spatial time series with nonstationary components that are both easily identifiable (annual periodicities) and difficult to view (trends and data anomalies). Since typical plots of single variables across time fail to show anything but the most fundamental components of this dataset, further efforts are needed to appreciate the more subtle elements in the data. Working with a fundamental principle that "the data display is the model" for the data, plots that show pertinent sources of variability in the data will be developed in an effort to better understand the nature of the dataset as a whole.
|
