|Click on any hyperlinked topic for an abstract|
|12:30 - 1:00||Registration, Posters, Tour sign-up|
|1:00 - 1:10||Opening remarks|
|1:10 - 2:00||Data Mining at IMS America|
|2:00 - 2:15||Breakout - Refreshments, Posters, Tour sign-up|
|2:15 - 3:05||Evaluation of Client/Server Configurations for Analytic Processing|
|3:05 - 3:20||Breakout - Refreshments, Posters, Tour sign-up|
|3:20 - 3:30||Open Forum|
|3:30 - 4:20||Advanced Applications of PROC REPORT|
|4:20 - 4:30||Wrap-up and Closing Remarks. Gather for tour|
|4:30 - 5:00||Optional Tour of IMS Data Center|
|Light refreshments will be served during breaks|
|A tour of the IMS America data center is available to all attendees. To go on this tour, which will last about a half hour, you must sign up during the registration time, or during the coffee breaks. The tour will begin shortly after the meeting ends.|
You are invited to join the speakers and the PhilaSUG Executive Committee for dinner at a nearby restaurant. The location will be announced at the meeting.
Data Mining at IMS America, How we turned a mountain of data into a few information-rich mole hills
IMS America, Inc.
This paper will illustrate how SAS/AF was used in a client-server environment using neural networking, SAS/Internet and other statistical techniques, to develop an application to allow pharmaceutical companies to detect brand switching, brand loyalty, and product trends in the pharmaceutical market. This data-mining system, designed as part of the IMS Xplorerb product, uses SAS components for data retrieval, data preparation, graphical user interface, and data visualization.
Over the past decade the IMS database has gone from megabytes to terabytes and is continuing to grow at an alarming rate. Until recently, clients who needed information from this database were restricted to various types of standard reports using business objects and other multi-dimensional analysis tools. Now, thanks to this data mining application, IMS clients can do their own data mining, turning a mountain of data into several information-rich molehills.
Our goal at IMS was to move from providing simple reporting services to our clients, to providing complex decision support. Using SAS software, we have been able to provide our customers with a powerful data mining tool that can give them a competitive advantage in the pharmaceutical industry.
Paul Kallukaran graduated with a Bachelors in Industrial and Production Engineering and a Masters of Science in Operations Research from Illinois Institute of Technology in Chicago. Since graduation, he has worked in the marketing research industry using his expertise in statistics, operations research, and artificial intelligence. Currently, he is a Manager of Research & Development in the Statistical Services Department at IMS America, a division of the Cognizant Corporation, and is pursuing a Master of Software Engineering at the Pennsylvania State University.
Jerry Kagan is an independent SAS consultant specializing in SAS/AF development. He has Bachelors degree in Management Science and an Associates degree in Computer Science from The Pennsylvania State University and has been using SAS software for the past eight years. Working primarily in the pharmaceutical and health care industries, Jerry has worked with companies including Wyeth-Ayerst Research, SmithKline Beecham, CoreStates Bank and The Prudential. He won a best paper award at SUGI 17 and was an invited speaker at SUGI 18.
Evaluation of Client/Server Configurations for Analytic Processing
Barry R. Cohen, Planning Data Systems, Inc.
I participated in a project to design and performance test a statistician's work station in a client/server environment. The mix included Oracle on HP/UNIX servers as the data warehouse, SAS Software for analytic processing, and both the HP/UNIX servers and Windows PC clients as candidate sites for the SAS program execution and data serving. The performance test results were presented in my SUGI21 paper. Much time was required in that paper to detail the hardware and software involved, and to simply present the performance results. Little time was available for critical analysis of the findings. I now build upon the information in that paper, taking a critical look at what the performance numbers tell us about setting up a client/server environment for data analysis with SAS Software. This includes examination of the real cost of moving data over a network at the time it is analyzed, and how that cost sometimes leads to avoiding the client/server set up in favor of one where all tasks are handled on one platform, be it a server platform or a PC platform. This paper should be of interest to people who are designing a client/server environment for analytical processing with SAS, where data set sizes range from tens of megabytes to perhaps several hundred megabytes, and are looking for some hard performance numbers and evaluation to guide them.
Barry Cohen is an Information Systems Consultant and President of Planning Data Systems, Inc., a consulting firm in Ardmore, Pennsylvania. His career spans 20 years of programming and systems development includes 17 years with SAS Software, providing services to a broad range of industries with a focus on the pharmaceutical industry. His most recent experiences have involved: the application of object-oriented programming techniques to pharmaceutical CANDA development through the SAS/AF and SCL, and the design and performance testing of various client server configurations for analytic processing. Barry is a co-founder and President of PhilaSUG, the Philadelphia Area SAS Users Group.
Advanced Applications of Proc Report
Ian Duling, SmithKline Beecham Pharmaceuticals.
This paper attempts to provide useful discoveries by the author in using the SAS Report Procedure, which provide solutions to specific problems faced in the production of data summary tables. The author has identified a pattern of predictable responses from the Report procedure to manipulations of the Group option and the Across option of the Define statement in order to create cross sectional tabulations. These innovations have made it possible for the author to respond to the majority of client report specifications through specialization in the use of this one procedure.
Using the Across option in particular, has presented challenges that require an understanding of internal processing that takes place as the procedure is executed. The effects of catagorical variable values on the Across option, particularly as the execution of this option is layered, can be managed through familiarity with the data and the use of other options such as Nozero. The ability to dynamically display data in column headers as part of the execution of the Across option is also a possibility.
Understanding how to reference the automatic variable _COL_ provides greater flexibility in the calculation and manipulation of columns within a cross tabulation. With the option to output a dataset (OUT=) and the SHOWALL option, the user can gain greater insight as to the effects of using the GROUP option in combination with the ACROSS option. This makes it easier to understand the effects of catagorical variables on the layout of tabular reports.
While the Group option requires an analysis variable on which to calculate a statistic, in order to perform the "consolidation into one row all observations from the data set that have unique combination of values for group variables", it is possible to satisfy that requirement and produce a table that gives the appearance of containing only character data.
The Report procedure can provide a means of quickly and easily customizing the look of a report using a data set that has been pre-processsed through the use of other procedures such as the Frequency procedure. This provides the capability to produce multiple summaries and statistics of the same data set within a one page report.
Ian Duling is a Clinical Analyst/Programmer at SmithKline Beecham Pharmaceuticals. For the past ten years, he has supported the reporting of clinical studies and electronic submissions to regulatory authorities, utilizing SAS software including SAS macros, various SAS procedures, and most recently SAS PH-Clinical. Ian received a BS degree in Business Administration from the University of Delaware in 1978.
Building PC Applications to Dynamically Access and Update Data on Multiple Platforms Using SAS AF and SAS Connect
Steve Rhoades, IMS America, Inc.
SAS Connect in combination with SAS AF takes advantage of client/server technologies to create easy to use point and click PC applications for end-users. This particular application uses a number of mainframe files as point and click reference files to build a mainframe batch maintenance record file. The reference files (selection lists) are built dynamically based on either a partial name or number. Behind the scenes all file opens, reads, displays, updates and closes are handled by the AF application with SAS Connect.
From the end-user's perspective, the application appears to be PC based. The only indication that it is not just a PC based system are the pop-up windows requesting User id and password. By pointing and clicking, various fields are populated for each maintenance record. Once the record is built and the 'Update File' button is selected, the record is added to the mainframe file and the build record process begins anew.
This poster will describe the process behind the scenes to accomplish this. This code was developed and tested using an MVS mainframe running SAS 6.08 and PC running SAS 6.12 under Windows 95.
The Six Ampersand Solution
John Gerlach, IMS America
Assume you want to reinstate macro variables representing criteria for an an analysis performed months ago. Using Dictionary tables made available through the SQL procedure, it is possible to collect macro variables used during the original analysis, and store them in a permanent SAS data set containing the name of each macro and its respective value. This paper explains how to reinstate those macro variables, which requires a six ampersand solution.
John Gerlach has been using the SAS System for over 10 years, specializing in the Macro Language and relational database management, and has worked mostly in the health industry, specifically epidemiology, pharmacology, and clinical trials. Currently, John is an independent contractor working on a large patient longitudinal data analysis project at IMS America. He has written and taught SAS seminars, as well.
Web-based SAS Output Viewer System
Shi-Tao Yeh, SmithKline Beecham Pharmaceuticals
This paper presents the SAS® macros to generate HTML-based files for publishing SAS output and SAS graphics onto Web pages. The viewer macro can create a SAS output viewer system to review and manage these SAS output Web pages. The discussion includes: · Preparing SAS list and graphics output, SAS macro to convert SAS list output into HTML code, SAS macro to convert SAS graphics into HTML code, and SAS macro to display Web-based SAS output.
The SAS products used in this paper are base SAS® and SAS/GRAPH®, with no limitation of operating systems.
Shi-Tao Yeh is a consultant at EDP Contract Services with assignments to SmithKline Beecham Pharmaceuticals. His areas of expertise are SAS Base, SAS/STAT, SAS/GRAPH, SAS/AF and SAS/FSP. He has a Ph.D. degree from the University of Pennsylvania and has been using SAS software for twenty years.
Last Update: 11:43PM 10/27/97