Hello, my name is James, and I’m an addict. I’m addicted to data science books, web courses, instructional videos, blogs, data science podcasts, predictive modeling competitions, and coding. This addiction takes up the majority of my mental energy. From the time that I wake up until I fall asleep (and all through my dreams), I’m generally thinking about data science concepts and coding. I’m going to bet that many of you are in a similar situation. If so, I’m sure that you have been as frustrated as I have been about the massive hole in the instructional data science market.
The market is overrun with data science books for Python, R, and Hadoop. These books provide an overview of data science and in-depth instructions on the various machine learning models, and they provide the associated development code for those particular programming languages. Although these books are great resources for data scientists, they do not offer direct programming instruction to the most popular programming language in the business community. SAS is used by 95% of Fortune 100 companies, and these companies are the leading employers of data scientists. There is an incredible opportunity to fill the need of professional data scientists for hands-on machine learning training with real-world examples.
The unfortunate reality for many SAS programmers is that we often do not have access to the latest and greatest SAS products. SAS Enterprise Miner, SAS Visual Analytics, SAS Forecast Server, and SAS Viya are all incredible products, but they are not universally available to all SAS programmers. It is essential that a data scientist who is working in a SAS environment be able to develop and implement machine learning models in any SAS environment. Even if data scientists have access to SAS Viya, it is incredibly beneficial for them to have a solid understanding of the programming code that drives the models that they develop in SAS Viya.
This book, End-to-End Data Science in SAS®, provides all SAS programmers insight into the models, methodology, and SAS coding required to develop machine learning models in any industry. It also serves as a reference for programmers of any language who either want to expand their knowledge base or who have just been hired into a data scientist position where SAS is the preferred language.
The goal of this book is to provide clear and practical explanations of the data science environment, machine learning techniques, and the SAS code necessary for the proper development and evaluation of these highly desired techniques. These explanations are demonstrated with real-world business applications across a variety of industries. All code and data sets are publicly available in a dedicated GitHub repository.
If you are interested in this book, then you (or most likely the organization that you work for) have SAS installed on your computer. However, not all SAS installations are created equal. Some programmers work in Base SAS (also called PC SAS). Others have a variety of SAS software available to them:
● SAS Enterprise Guide
● SAS Enterprise Miner
● Visual Analytics
● SAS Studio
● SAS Viya
This list is just a sample of the many SAS products available. In addition to these products, there are several software components that SAS offers:
● SAS/ACCESS software
● SAS/ETS software
● SAS/IML software
● SAS ODS Graphics Editor
● SAS/OR software
● SAS/STAT software
Your company’s IT department generally dictates the SAS products that you have and the software components that are available to you. If you desperately want SAS Viya or SAS/ETS, you will often have to “fight the power” to get it. I sincerely hope that you can access one or many of these SAS products because they are awesome, and they will make your life as a data scientist much easier and much more productive. However, if you are like me and you have to develop predictive models without the benefit of all the toys that SAS has to offer, then this book is what you have been waiting for.
The minimum requirement for the majority of procedures detailed in this book is SAS 9.2 with SAS/STAT installed. This requirement should cover most SAS users. With this minimum requirement, we will be able to develop:
● Linear regressions
● Logistic regressions
● Clustering
● Decision trees
Some of the more advanced procedures will require SAS Enterprise Miner to be installed. These procedures will include:
● PROC HPFOREST
● PROC TREEBOOST
● PROC SVM
Don’t panic! Just because you are limited to Base SAS with SAS/STAT installed does not mean that we are limited to the predefined procedures that come with SAS/STAT. We are limited only by our ingenuity. I will include methods of how to perform some of these more advanced procedures by developing them from scratch, using only the procedures available in SAS/STAT. The code to create these procedures from scratch can get pretty complicated, but I’ll provide step-by-step explanations and all code will be available in the repository.
If you only have experience working with Microsoft Excel spreadsheets, then don’t worry; we can get you up and running in SAS. However, if you’ve never worked with data in any capacity, including Excel spreadsheets, then maybe this book isn’t for you.
I will assume that you have some experience writing basic formulas. For example, in Excel:
=AVERAGE(B12: B32)
=SUM(A1: A14)
I will also assume that you can perform basic computer commands such as create a new file, save a file, open a file, and so on.
As long as you have met these minimum requirements and have a desire to learn awesome new skills that are guaranteed to increase your value to any organization, then we are good to go.
|
|
This icon indicates a warning or a difficult subject |
|
|
This icon indicates a decision that the data scientist has to make |
All data used for examples in this book have been accessed through public-use data repositories. These repositories provide a wide range of data sets that span across a variety of disciplines. The code that is demonstrated in the book is primarily created by the author. Any code that has been shown that was not created by the author has been cited with the appropriate credit to the developer. There is a fantastic community of SAS developers who openly share their code with the world. I strongly encourage you to search the web and connect with these communities on the SAS website, GitHub, Stack Overflow, and several other communities.
You can access the example code and data for this book by visiting the GitHub repository at https://github.com/Gearhj/End-to-End-Data-Science.
This book is compatible with SAS University Edition. If you are using SAS University Edition, then begin here: https://support.sas.com/ue-data.
Do you have questions about a SAS Press book that you are reading? Contact us at [email protected].
SAS Press books are written by SAS Users for SAS Users. Please visit sas.com/books to sign up to request information on how to become a SAS Press author.
We welcome your participation in the development of new books and your feedback on SAS Press books that you are using. Please visit sas.com/books to sign up to review a book.
Learn about new books and exclusive discounts. Sign up for our new books mailing list today at https://support.sas.com/en/books/subscribe-books.html.
Learn more about this author by visiting his author page at http://support.sas.com/gearheart. There you can download free book excerpts, access example code and data, read the latest reviews, get updates, and more.
I would like to thank SAS for allowing me the opportunity to publish this book through SAS Press. It is an honor to be part of the SAS author community. I would also like to give special thanks to my editor Suzanne Morgen for her patience and her incredible eye for detail.
Additional thanks are given to all of the reviewers who painstakingly reviewed raw chapters and provided needed critical feedback:
● Christopher Battiston
● David Ghan
● Funda Gunes
● Sunil Gupta
● Mark Jordan
● Robin Langford
● Premkumar Varma
I would still be a poor musician in Warren, Ohio if it were not for Lorin Ranbom, Mitali Ghatak and the Health Services Research Section at the Ohio Department of Jobs and Family Services who took a chance on hiring me right out of college. Thanks to Dan Hecht, Donna Bush, Lora Summers, Kendy Markmen, Dave Dorsky, Hope McGonigle, Tracy Cloud and Eric Edwards for introducing me to SAS and providing me all of the support that I needed.
A continuous thank you to the best SAS programmer that I ever met, Matt Bates. I cannot count how many times that I bothered Matt about how to perform tasks from the mundane to the insanely complex. He has always come through for me and I am eternally grateful.
A huge thank you to Mihai Gavril, Vijay Narapsetty, Dan Gedrich, and Jay Das for providing a work environment that is collaborative, innovative and focused on excellence. I couldn’t ask for a better team.
Thank you to my parents Richard and Lou Ann Gearheart for all of their love and support and my brothers Scott, Glen, Kenny, George, and Tod. You have always been there for me.
My beautiful son, Jacob Gearheart, is the pride of my life and I am amazed every day with your wit, charm, intelligence, and kindness. I pray that I will be more like you every day.
And of course, the love of my life, Tanya Gearheart. I must have written you a thousand love letters and none of them have come close to how much that I love you and need you in my life. You are my entire world and the only reason that I wake up every day is to spend more time with you. Always, madly.