SAS

General Information

Short Description: SAS is an integrated collection of quantitative analysis,  and data management and warehousing software. It is used both in academic research,  academic administration and in private industry (across disciplines).

Long Description:  SAS, originally Statistical Analysis System, is an integrated system of software products.  The typical installation at Boston College includes Base SAS, SAS/GRAPH, SAS/ETS, SAS/Stat, and SAS/ASSIST.  With these modules, you can perform many data management and statistical analysis tasks.

Base SAS includes a fourth-generation programming language; ready-to-use programs for data  manipulation, information storage and retrieval, descriptive statistics, and report writing; and a macro facility.  You can manage data with case selection, file reshaping, and creating derived data. Statistical analysis tasks  that can be performed with the base package include the generation of descriptive statistics, bivariate statistics, prediction of numerical outcomes, and prediction of identifying groups.

With SAS/GRAPH software you can create charts and graphs.  Available charts include vertical and horizontal bar, pie, donut, subgrouped pie and donut, star and block, and many more.  Plots include scatter, line, area, bubble, multiple axis and overlay. Data also  can be displayed within a three-dimensional coordinate system using response surfaces or scatter plots.

SAS/ETS software offers a broad array of time series, forecasting and econometric techniques.

SAS/STAT allows researchers to perform analysis of variance, regression, categorical data analysis, multivariate analysis,  survival analysis, cluster analysis, psychometric analysis, nonparametric analysis, multiple imputation for missing values, and other statistical analysis procedures that are not  covered in Base SAS.

SAS is predominantly a syntax based language but SAS/ASSIST provides a task-oriented visual interface.

Category: Quantitative Data Analysis

Programming Language:  SAS Syntax Language.

Installation Information

Version:  9.2   Version Date:  2009

Technical support:  Please contact Research Services (researchservices@bc.edu).

Terms of Use: Available for teaching and research on BC computers and home computers of BC faculty and staff (Note: one home computer only), and the computer labs in the departments listed below.

Distribution: For installation on a BC-owned computer, please contact your local TC to schedule a time.  Faculty and staff members who have SAS installed on their BC-owned desktop/laptop computer may install SAS on one of their home computers.  Please contact researchservices@bc.edu for information on installing SAS on your home computer.

License Information

Number of Licenses:  Site License.

Type of License:  Individual licenses (Windows, Scorpio Linux Cluster). 

Installed Sites: Graduate School of Social Work (GSSW), Connell School of Nursing (CSON), Carroll School of Management (CSOM), Lynch School of  Education (LSOE), Gasson 09, and the CTRC, BC owned computers.

Installed Platforms: Windows, Linux (scorpio.bc.edu). 

Usage Information

On the Scorpio Linux Cluster:

First load the sas module (module add sas).  If  you use SAS often, you may want to add sas to the module line in your .tcshrc. Then it will be done for you everytime you login.

1.  For short usage of sas (10 minutes or less)

sas filename

2. For longer sas runs, please use the queues.  When you submit a job to the queues, it gets scheduled to run on another node of the cluster. You will get a dedicated processor and memory to yourself.

To use the queues, you first need to create a command file.  I will name it sas.pbs, but you can name it anything you want.  The file should contain:

#!/bin/tcsh
#PBS -l mem=1gb,nodes=sas+1:ppn=1,walltime=10:00:00
#PBS -m abe

module add sas

cd directory-where-your-sas-input-file-is
sas filename

You are requesting 1 GB of memory, and 10 hours of time.  Currently, the maximum time for SAS is 24 hours.   If you need more, please let us know (researchservices@bc.edu).

Note:  For large data sets use the following instead of "sas filename"

   sas filename -memsize 2048m -work /scratch/user-name

Now, submit it to the queues

module add torque
qsub sas.pbs

For more information on the queues, see the Torque (PBS) User Guide.

Vendor/Developer:  SAS Institute, Inc.

Vendor Website: www.sas.com

User Guide: Documentation is available at:
    support.sas.com/documentation/onlinedoc/91pdf/index_913.html