A time ago the Western Power’s System Forecasting team, where I now belong in, initiated an upgrade of the software used for data analysis and forecasting. Several options were considered, including SPSS Modeller and a few industry specific options, but the final decision was to use SAS.

One reason for this desirable shift (R was not considered a suitable option) was to make use of advanced data techniques and models to quickly build reliable and accurate forecasts to improve business decisions. SAS Institute claims their current technologies can convert “transactional meter and operational systems data into a forecast-ready format”, which is exactly what the team required. However, for a corporation as big as Western Power, it may take another few years before all SAS components required to achieve this goal can be properly integrated and optimised. This time can be used to learn how to use SAS most efficiently and effectively.

There are many on-line resources with plenty of code available to help one learn SAS for free, e.g. UCLA’s SAS learning modules and Robert Allison’s SAS/Graph examples, but books published by SAS Press may remain the best learning option for beginners. It is the purpose of this post to list the books that are likely to be useful for the type of tasks the team performs now and will be expected to perform in the near future.

The list is not meant to be exhaustive, and in particular it does not include some older books on time series forecasting. More recent books on time series analysis, forecasting and predictive modelling are in as most relevant. The general programming books (functions, macros) are useful for developing a general understanding of and intuition about the particularities of the SAS programming language. The book on statistical programming with SAS/IML (Interactive Matrix Language) contains many examples of using R from within SAS environment. The two books on SQL will help learn how to interact with the company’s databases. There is also a book devoted solely to reports, which I think should not be ignored, and a book on techniques useful for cleaning data.

The list includes 12 items, and the first few have already been purchased:

  1. Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications by Kattamuri S. Sarma, 2nd ed., 2013.

  2. Applied Data Mining for Forecasting Using SAS by Tim Rey, Arthur Kordon, and Chip Wells, 2012

  3. SAS functions by example by Ron Code, 2nd ed., 2010

  4. SAS for Forecasting Time Series by John C. Brocklebank and David A. Dickey, 2nd ed., 2003

  5. Practical Time Series Analysis Using SAS by Anders Milhoj, 2013

  6. PROC REPORT by Example: Techniques for Building Professional Reports Using SAS, by Lisa Fine, 2013

  7. PROC SQL: Beyond the Basics Using SAS by Kirk Paul Lafler, 2nd ed., 2013

  8. SAS Programming for Enterprise Guide Users by Neil Constable, 2nd ed., 2010

  9. Statistical Programming with SAS/IML Software by Rick Wicklin, 2010

  10. Cody’s Data Cleaning Techniques Using SAS by Ron Cody, 2nd ed., 2008

  11. PROC SQL by Example: Using SQL within SAS by Howard Schreier, 2008

  12. SAS Macro Programming Made Easy by Michele Burlew, 2nd ed., 2007

Time series models: VAR model definition

The Vector AutoRegression (VAR) family of models has been widely used for modelling and forecasting since the early 1980s. A VAR model is...… Continue reading

Time series models: ARIMA model definition

Published on September 30, 2015

Types of forecast: ex ante vs ex post

Published on September 06, 2015