tips and tricks - analytics, business intelligence and data … group presentations... · tips and...
TRANSCRIPT
Tips and Tricks from the SAS Global Forum 2012
MONSUG - May 25th 2012
By Eric Lacombe, Director, Professional Services
CLEARGOALS Company
May 25th 2012
Agenda Overview
Agenda
Copyright 2012 CLEARGOALS 3 Confidential
Selection Process
Goal of the Presentation
Insights from the 10 Selected Papers
Conclusion
Selection Process
Selection Process
411 papers in pdf format – read twice the abstracts
Down to 87 papers – read twice (and more) the detailed papers
Down to 30 down to 13 down to 11 down to 10
A lot of interesting material
Difficult to make a choice
Tried to pick quick win paper content
Tried to reach a broad audience with various paper subjects
Copyright 2012 CLEARGOALS 5 Confidential
Goal of the Presentation
Goal
Provide an overview of the paper
And
Provide a few tricks from the paper
Copyright 2012 CLEARGOALS 7 Confidential
Insights from the 10 Selected Papers
Insights – Paper #1
Section: Coders’ Corner
Paper 047-2012
Working the System: Our Best SAS® Options
Patrick Thornton, SRI International, Menlo Park, CA
Iuliana Barbalau, Adecco, Pleasanton, CA
• http://support.sas.com/resources/papers/proceedings12/047-2012.pdf
Copyright 2012 CLEARGOALS Confidential 9
Insights – Paper #1
• Discussion on beneficial and interesting options
Copyright 2012 CLEARGOALS Confidential 10
Insights – Paper #1
• Option : PDFSECURITY
Copyright 2012 CLEARGOALS Confidential 11
Insights – Paper #1
• Options : MCOMPILE, MCOMPILENOTE (New in SAS 9.2)
Copyright 2012 CLEARGOALS Confidential 12
Insights – Paper #2
Section: Programming: Foundations and Fundamentals
Paper 241-2012
A Survey of Some of the Most Useful SAS® Functions
Ron Cody, Camp Verde, Texas
• http://support.sas.com/resources/papers/proceedings12/241-2012.pdf
Copyright 2012 CLEARGOALS Confidential 13
Insights – Paper #2
• Covered Functions:
– LENGTHN, LENGTHC – MISSING, CALL MISSING ROUTINE – INPUT, PUT – FIND, FINDC – COMPRESS – SUBSTR – SCAN – UPCASE, LOWCASE, PROPCASE – TRANWRD – SPEDIS – TRIMN, STRIP – NOTALPHA, NOTDIGIT, NOTALNUM – CATS, CATX – COUNT, COUNTC – MDY, MONTH, WEEKDAY, DAY, YEAR, YRDIF – ARRAY – N,NMISS, SUM, MEAN – SMALLEST, LARGEST – LAG – CALL SORTN ROUTINE
Copyright 2012 CLEARGOALS Confidential 14
Insights – Paper #2
• COMPRESS function
Copyright 2012 CLEARGOALS Confidential 15
Insights – Paper #2
• Example – Program
– Output
Copyright 2012 CLEARGOALS Confidential 16
Insights – Paper #2
• PROPCASE function – stands for proper case
– capitalizes the first character in each "word" and sets the remaining letters to lowercase.
Program
Output
Copyright 2012 CLEARGOALS Confidential 17
Insights – Paper #3
Section: Programming: Beyond the Basics
Paper 227-2012
Executing a PROC from a DATA Step
Jason Secosky, SAS Institute Inc., Cary, NC
• http://support.sas.com/resources/papers/proceedings12/227-2012.pdf
Copyright 2012 CLEARGOALS Confidential 18
Insights – Paper #3
• Detailed paper on how to execute a PROC from a DATA step
• In SAS 9.2, RUN_MACRO enables DATA step code to immediately execute a macro an act on the result
• In SAS 9.3, SAS introduced the DOSUBL function
– Experimental function that executes SAS code directly from a DATA step
– Unlike RUN_MACRO, DOSUBL can be called directly from a DATA step without the need for a user-written function
– Character value required
Copyright 2012 CLEARGOALS Confidential 19
Insights – Paper #3
• DOSUBL function – Code example
– You can use the DATA step OPEN and FETCH functions to open and read the resulting dataset WORK.REGIONS.
Copyright 2012 CLEARGOALS Confidential 20
Insights – Paper #4
Section: SAS Enterprise Guide® Implementation and Usage
Paper 301-2012
Productivity Tips for SAS® Enterprise Guide® Users
Jennifer First and Steven First, Systems Seminar Consultants, Madison, WI,
United States
• http://support.sas.com/resources/papers/proceedings12/301-2012.pdf
Copyright 2012 CLEARGOALS Confidential 21
Insights – Paper #4
• The paper covers
– EG OPTIONS • PROJECT VIEWS
• RESULTS
• DATA AND QUERY OPTIONS
• EDITOR OPTIONS
• LAYOUT OPTIONS
– ORGANIZING SAS PROCESSES • DOCUMENTATION
• MULTIPLE PROCESS FLOWS
• PROCESS FLOW ORGANIZATION
– LEVERAGING POINT AND CLICK FUNCTIONALITY • QUERY BUILDER
Copyright 2012 CLEARGOALS Confidential 22
Insights – Paper #4
• ORGANIZING SAS PROCESSES – DOCUMENTATION
• The Note
• A text document where a user can type details
• It will appear as a project node in the Process Flow and the Project Tree
• A best practice in EG is to create a Note linked to each task used in the Project
Copyright 2012 CLEARGOALS Confidential 23
Insights – Paper #4
• Example of a note associated with a task that documents the task
Copyright 2012 CLEARGOALS Confidential 24
Insights – Paper #4
• LEVERAGING POINT AND CLICK FUNCTIONALITY
– QUERY BUILDER
• JOINING TABLE
– Up to 32 tables
– The user doesn’t need to worry about correctly spelling variable names correctly
– Saves time
– Grab the generated code to modify it later
Copyright 2012 CLEARGOALS Confidential 25
Insights – Paper #4
• Example of code generated by the Query Builder
Copyright 2012 CLEARGOALS Confidential 26
Insights – Paper #4
• LEVERAGING POINT AND CLICK FUNCTIONALITY
– Easy to create graphs
• Creating graphs from SAS Coding can often be time consuming and cumbersome
• In the past, many SAS users have even exported their results to Excel to create their graphs
– Easy to create processes that are difficult to code
• Leverage the task
• For example, the Summary Tables Task (generated PROC TABULATE code)
Copyright 2012 CLEARGOALS Confidential 27
Insights – Paper #4
• Example of a Summary Tables Task (PROC TABULATE)
Copyright 2012 CLEARGOALS Confidential 28
Insights – Paper #5
Section: Posters
Paper 217-2012
Wake up your data with Graph’n’Go
Christopher Battiston, Hospital for Sick Children, Toronto, Ontario, Canada
• http://support.sas.com/resources/papers/proceedings12/217-2012.pdf
Copyright 2012 CLEARGOALS Confidential 29
Insights – Paper #5
• Unknown feature to create graphs and visualize data
• Easy to use, powerful, flexible
• Can output to SAS and modify the code
• Available since SAS 8
• Requires SAS/Graph
• To start the tool – Go to Solutions and then Reporting
Copyright 2012 CLEARGOALS Confidential 30
Insights – Paper #5
• How-to
– Select data
Copyright 2012 CLEARGOALS Confidential 31
Insights – Paper #5
• How-to
– Change some of the characteristics of the graph
Copyright 2012 CLEARGOALS Confidential 32
Insights – Paper #5
• How-to
– The completed set of graphs
Copyright 2012 CLEARGOALS Confidential 33
Insights – Paper #5
• How-to
– Export the code (PROC GCHART)
Copyright 2012 CLEARGOALS Confidential 34
Insights – Paper #6
Section: Hands-on Workshops
Paper 153-2012
Quick Results with SAS® ODS Graphics Designer
Sanjay Matange, SAS Institute Inc., Cary, NC
• http://support.sas.com/resources/papers/proceedings12/153-2012.pdf
Copyright 2012 CLEARGOALS Confidential 35
Insights – Paper #6
• Don’t need to know the SG Procedures (ODS Graphics Procedures)
• Interactive ‘drag-and-drop’ process
• Released with the 3rd maintenance release for SAS 9.2
• Part of the SAS/GRAPH product
• Starting with SAS 9.3, the designer became available as part of Base SAS
• The designer is based on Graph Template Language (GTL)
– Used to create analytical graphs produced by the SAS analytical procedures
• Can save the graph to a file or a gallery
• Possibility to run the designer graphs in batch
Copyright 2012 CLEARGOALS Confidential 36
Insights – Paper #6
• GETTING STARTED 1. Submit the %sgdesign macro in the Program Editor
2. With SAS 9.3, you can start the designer from the Tools -> ODS Graphics Designer menu entry
Copyright 2012 CLEARGOALS Confidential 37
Insights – Paper #6
• THE DESIGNER APPLICATION INTERFACE
Copyright 2012 CLEARGOALS Confidential 38
Insights – Paper #6
• Distribution of Hamilton Scores in 14 easy steps
Copyright 2012 CLEARGOALS Confidential 39
Insights – Paper #6
• Survival Plot
Copyright 2012 CLEARGOALS Confidential 40
Insights – Paper #6
• Classification Panel
Copyright 2012 CLEARGOALS Confidential 41
Insights – Paper #7
Section: Programming: Foundations and Fundamentals
Paper 259-2012
Graphing Made Easy with SG Procedures
Lora D. Delwiche, University of California, Davis, CA
Susan J. Slaughter, Avocet Solutions, Davis, CA
• http://support.sas.com/resources/papers/proceedings12/259-2012.pdf
Copyright 2012 CLEARGOALS Confidential 42
Insights – Paper #7
• New in SAS 9.2 as a part of SAS/GRAPH
• Included in Base SAS in SAS 9.3
• New way of generating graphs using SAS
• 3 basic Statistical Graphics (SG) procedures (SGPLOT, SGPANEL, SGSCATTER)
• SGPLOT
– Can produce 18 different types of graphs (bar charts, histograms and density plots, scatter plots)
• The paper gives all the possible options that can be used with the procedures
• Possibility to save the ODS GRAPHICS output and access individual graphs
Copyright 2012 CLEARGOALS Confidential 43
Insights – Paper #7
• PROC SGPLOT Examples
Copyright 2012 CLEARGOALS Confidential 44
Insights – Paper #7
• PROC SGPLOT Examples
Copyright 2012 CLEARGOALS Confidential 45
Insights – Paper #7
• PROC SGPLOT Examples
Copyright 2012 CLEARGOALS Confidential 46
Insights – Paper #7
• PROC SGPLOT Examples
Copyright 2012 CLEARGOALS Confidential 47
Insights – Paper #8
Section: Statistics and Data Analysis
Paper 313-2012
Look Out: After SAS/STAT® 9.3 Comes SAS/STAT 12.1!
Maura Stokes, Fang Chen, Yang Yuan, and Weijie Cai
SAS Institute, Inc. Cary NC
• http://support.sas.com/resources/papers/proceedings12/313-2012.pdf
Copyright 2012 CLEARGOALS Confidential 48
Insights – Paper #8
• This paper reviews the highlights of the new release and illustrates them with practical examples
• New release-numbering scheme • Expect more frequent releases of SAS/STAT software (each 12-18 months) • SAS/STAT 9.3 became available in 2011
– Introduces the experimental FMM procedure which fits statistical models to date where the distribution of the response is a finite mixture of univariate distributions
– The MI procedure added the FCS statement, which specifies a multivariate imputation by fully conditional specification (FCS) methods
– The NLIN procedure was updated with features for diagnosing the nonlinear model fit. The SURVEYPHREG procedure became production and now handles time-dependent covariates
– The MCMC procedure added a RANDOM statement, which simplifies the specification of hierarchical random-effects models and significantly reduces simulation time while improving convergence
Copyright 2012 CLEARGOALS Confidential 49
Insights – Paper #8
• The upcoming SAS/STAT 12.1 emphasizes modern regression methods
– The new QUANTSELECT procedure for quantile regression model selection works similarly to the GLMSELECT procedure
– The new QUANTLIFE procedure performs quantile regression for censored data
– The new ADAPTIVEREG procedure provides flexible regression model for high-dimensional data
– In addition, epidemiologists will benefit from the new STDRATE procedure, which computes direct and indirect standardized rates and risks for study populations
– The FMM procedure for finite mixture models becomes production, and Bayesian analysis capabilities are also updated
Copyright 2012 CLEARGOALS Confidential 50
Insights – Paper #8
• A number of existing procedures have also had important updates
– WEIGHT statement in PROC LIFETEST
– case-level (observation-level) residual diagnostics with latent variables in PROC CALIS
– partial R-square for relative importance of parameters in PROC LOGISTIC
– Miettinen-Nurminen confidence limits for the difference of proportions in PROC FREQ
– Poisson sampling in PROC SURVEYSELECT
– group sequential design with nonbinding acceptance boundary in the SEQDESIGN and SEQTEST procedures
– post-stratification estimation in the SURVEYMEANS procedure
– REF= option added to the CLASS statement for GLM, MIXED, GLIMMIX, and ORTHOREG procedures
Copyright 2012 CLEARGOALS Confidential 51
Insights – Paper #9
Section: Systems Architecture and Administration
Paper 366-2012
Logging 101: Leveraging the SAS Enhanced Logging Facility
Margaret Crevar, SAS Institute Inc., Cary, NC, USA
Tony Brown, SAS Institute Inc., Dallas, TX USA
• http://support.sas.com/resources/papers/proceedings12/366-2012.pdf
Copyright 2012 CLEARGOALS Confidential 52
Insights – Paper #9
• Beginning with SAS 9.2, a new, enhanced logging facility was incorporated into Base SAS
• This facility provides the ability to finely tune the capturing of log and audit information about SAS activity
• Since the captured data is stored in external files, SAS programmers can use SAS 4GL language statements in custom-coded programs to report on this data
• The enhanced logging features classify messages according to audit, performance, administration, and application namespaces, as well as by diagnostic levels such as trace, information, error, warning, and fatal
• Through the use of appenders, administrators can direct messages to customized output destinations such as the Windows Event Viewer, UNIX system logs, the z/OS system logger, or an operator’s console
Copyright 2012 CLEARGOALS Confidential 53
Insights – Paper #9
• Can gather statistics from Enterprise BI Applications or ad hoc and batch processes
• For the SAS batch jobs
– The collection of information from these processes requires that the logging facility be turned on globally for all SAS processes
– Can be done with a global AUTOEXEC.SAS file that is used by all SAS sessions. You simply add the logging facility autocall macros to the top of the ad hoc or batch SAS program code for each job
– The MAUTOSOURCE system option must be set and the %LOG4SAS autocall macro must be invoked before any other logging facility autocall macros are invoked
– Placing this autocall macro within the job automatically creates the appropriate logging information
Copyright 2012 CLEARGOALS Confidential 54
Insights – Paper #9
• The SAS Audit, Performance and Measurement package acquires the data collected by the enhanced logging facility and uses it to create predefined sample reports that can help SAS and IT administrators understand how SAS is being used on their various computer hardware systems
• Reports such as Most Heavily Used Datasets/Directories, Top 10 Reports/OLAP Cubes/Users, and SAS PROC Usage are just examples of the types of reporting that can be done
• The information can easily be expanded to cover the complete set of computer resources being used by SAS
• The information can even be used as input to internal chargeback systems
Copyright 2012 CLEARGOALS Confidential 55
Insights – Paper #9
• A sample Web application included with the SAS Audit, Performance and Measurement package reports on the SAS Enterprise BI content (such as reports, data, and projects) that are accessed by individual SAS Enterprise Guide processes. It also reports on the resulting computer resource consumption
• Some case study examples are available in the paper
– Available disk space in the SAS WORK file system
– SAS Metadata Server
Copyright 2012 CLEARGOALS Confidential 56
Insights – Paper #9
• The SAS Audit, Performance and Measurement Package with the Usage Reports Highlighted
Copyright 2012 CLEARGOALS Confidential 57
Insights – Paper #9
• SAS 9.3 Additions to the SAS Audit, Performance and Measurement Package
Copyright 2012 CLEARGOALS Confidential 58
Insights – Paper #10
Section: Programming: Data Management
Paper 115-2012
What’s New in SAS/ACCESS®
and Process Improvements to Apply to Your DBMS
Howard Plemmons, SAS Institute Inc., Cary, NC
• http://support.sas.com/resources/papers/proceedings12/115-2012.pdf
Copyright 2012 CLEARGOALS Confidential 59
Insights – Paper #10
• This paper provides an overview of major new features in SAS/ACCESS and looks at process improvements that can help with DBMS performance
• The paper contains enhancements that are in SAS 9.3M1 and what is being developed and slated for SAS 9.3M2
• Supported Interfaces and Platforms
Copyright 2012 CLEARGOALS Confidential 60
Insights – Paper #10
• For SAS 9.3M1, SAS has provided additional parse/textualization processes to support implicit pass-through SQL. These components were added to PROC SQL to enable push-down of processing into the DBMS
• The DBIDIRECTEXEC option must be set
Copyright 2012 CLEARGOALS Confidential 61
Insights – Paper #10
• SAS/ACCESS ENGINE READ PERFORMANCE - pipeline data reads
– This feature is slated for SAS 9.3M2
– The Pipeline read feature is designed to give the user an automated way to gain read performance without having to tweak and experiment with the READBUFF option
– The more you process data in SAS, the faster it can be pipelined into the SAS/ACCESS read process
Copyright 2012 CLEARGOALS Confidential 62
Insights – Paper #10
• Performance With and Without Pipeline Read
Copyright 2012 CLEARGOALS Confidential 63
Conclusion
Conclusion
• Feel free to present a paper or a book that you have read or a presentation or a video that you have seen
• Feel free to contact me for additional information and questions on the presentation
LinkedIn: http://ca.linkedin.com/in/ericlacombe
Copyright 2012 CLEARGOALS Confidential 65
THE END
Copyright 2012 CLEARGOALS 66 Confidential
Thank You