Chapter Contents |
Previous |
Next |

Getting Started |

To explore data, you can

- identify observations in plots
- brush observations in linked scatter plots, histograms, box plots, line plots, contour plots, and three-dimensional rotating plots
- exclude observations from graphs and analyses
- search, sort, and edit data
- transform variables
- color observations based on the value of a variable

To analyze distributions, you can

- compute descriptive statistics
- create quantile-quantile plots
- create mosaic plots of cross-classified data
- fit parametric (normal, lognormal, exponential, Weibull) and kernel density estimates for distributions
- fit parametric and empirical cumulative distribution functions
- test hypotheses of completely specified (known parameters) or specific (unknown parameters) parametric distributions based on Kolmogorov's D statistic

To analyze relationships between a response variable and a set of explanatory variables, you can

- fit curves with polynomials, kernels, and smoothing splines
- fit curves with nonparametric local polynomial smoothers using either a fixed bandwidth or loess smoothing
- add confidence bands for mean and predicted values
- fit surfaces with polynomials, kernels, and smoothing splines
- create residual and leverage plots
- fit the general linear model, including classification effects for analysis of variance and analysis of covariance
- fit the generalized linear model, including logistic regression, Poisson regression, and other models

To analyze relationships between variables, you can

- calculate correlation matrices and scatter plot matrices with confidence ellipses for relationships among pairs of variables
- reduce dimensionality of interval variables with principal component analysis
- examine relationships between two sets of interval variables with canonical correlation analysis and maximum redundancy analysis
- examine relationships between a nominal variable and a set of interval variables with canonical discriminant analysis

In addition, you can

- process data by groups
- process multiple data sets
- store option settings to customize SAS/INSIGHT operation
- store results as SAS data sets, SAS/GRAPH catalogs, and text files
- record and submit SAS/INSIGHT statements
- obtain context-sensitive help

Finally, because it is a part of the SAS System, you can use SAS/INSIGHT software to explore results from any SAS procedure. Conversely, you can use any SAS procedure to analyze results from SAS/INSIGHT software.

Chapter Contents |
Previous |
Next |
Top |

Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.