|
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 - #305477 |
|
Title:
|
Exploratory Data Analysis of Meteorological Data Using SAS Stat Studio
|
|
Author(s):
|
Frederick Wicklin*+ and Yun Chen
|
|
Companies:
|
SAS Institute, Inc. and North Carolina State University
|
|
Address:
|
SAS Campus Drive, Cary, NC, 27513,
|
|
Keywords:
|
meteorological data ; exploratory data analysis ; statistical software ; dynamically-linked graphics ; SAS STAT studio ; semivariogram
|
|
Abstract:
|
We explored the geographic and atmospheric variables in NASA meteorological data. We discovered inconsistent pressure readings, indicating a possible systematic measurement error. We discovered geographic categories for which the atmospheric variables are similar. We adjusted the atmospheric variables to account for the geographic effects, then formulated models that explain the relationship between ground-based temperature measurement and satellite-based measurements. Similarly, we formulated models to explore the seasonal and long-term variation of spatial averages of the atmospheric variables. Lastly, we constructed empirical semivariograms to visualize the spatial dependence of temporal averages of the atmospheric variables. We analyzed this data using Stat Studio, a new statistical software product from SAS that provides dynamically-linked graphics for exploring multivariate data.
|
