Python for Marketing Research and Analytics
Авторы: Jason S. Schwarz, Chris Chapman, Elea McDonnell Feit (2020)
We are here to help you learn Python for marketing research and analytics. Python is a great choice for marketing analysts. It offers advanced capabilities for fitting statistical models. It is extensible and is able to process data from many different systems, in a variety of forms, for both small and large datasets. The Python ecosystem includes a vast range of established and emerging statistical methods as well as visualization techniques. Yet its use in marketing lags other fields such as econometrics, bioinformatics, and computer science. With your help, we hope to change that! This book is designed for two audiences: practicing marketing researchers and analysts who want to learn Python and students or researchers from other fields who want to review selected marketing topics in a Python context. What are the prerequisites? Simply that you are interested in Python for marketing, are conceptually familiar with basic statistical models such as linear regression, and are willing to engage in hands-on learning. This book will be particularly helpful to analysts who have some degree of programming experience and wish to learn Python. In Chap. 1, we describe additional reasons to use Python (and a few reasons perhaps not to use Python). The hands-on part is important. We teach concepts gradually in a sequence across the first seven chapters and ask you to type our examples as you work; this book is not a cookbook style reference. We spend some time (as little as possible) in Part I on the basics of the Python language and then turn in Part II to applied, real-world marketing analytics problems. Part III presents a few advanced marketing topics. Every chapter shows off the power of Python, and we hope each one will teach you something new and interesting. Specific features of this book are: • It is organized around marketing research tasks. Instead of generic examples, we put methods into the context of marketing questions. • We presume only basic statistics knowledge and use a minimum of mathematics. This book is designed to be approachable for practitioners and does not dwell on equations or mathematical details of statistical models (although we give references to those texts). • This is a didactic book that explains statistical concepts and the Python code. We want you to understand what we are doing and learn how to avoid common problems in both statistics and Python. We intend the book to be readable and to fulfill a different need than references and cookbooks available elsewhere. • The applied chapters demonstrate progressive model building. We do not present “the answer” but instead show how an analyst might realistically conduct analyses in successive steps where multiple models are compared for statistical strength and practical utility. • The chapters include visualization as a part of core analyses. We do not regard visualization as a standalone topic; rather, we believe it is an integral part of data exploration and model building. • Most of the analyses use simulated data, which provides practice in the Python language along with additional insight into the structure of marketing data. If you are inclined, you can change the data simulation and see how the statistical models are affected. • Where appropriate, we call out more advanced material on programming or models so that you may either skip it or read it, as you find appropriate. These sections are indicated by ∗ in their titles (such as This is an advanced section∗). What do we not cover? For one, this book teaches Python for marketing and does not teach marketing research in itself. We discuss many marketing topics but omit others that would simply repeat the analytic methods in Python. As noted above, we approach statistical models from a conceptual point of view and skip the mathematics. A few specialized topics have been omitted due to complexity and space; these include customer lifetime value models and econometric time series models. v vi Preface Overall, we believe the analyses here represent a great sample of marketing research and analytics practice. If you learn to perform these, you will be well equipped to apply Python in many areas of marketing. For another, this book teaches Python for marketing and not all of the complexity or subtlety of the Python programming language or of programming generally. We present the basics of Python programming that are needed to successfully analyze marketing data, but there is much more to Python than we present here. Companion Book: R for Marketing Research and Analytics This book is closely related to R for Marketing Research and Analytics (Chapman and Feit 2019) and shares many datasets and sections of didactic explanation of methods with that R text. In some ways, this Python book and the R text are mutual “translations” of one another from R and Python, respectively. This was a deliberate choice that we hope will make it easy for readers to move between Python and R. If you understand a method in one language, the companion book will demonstrate how to perform a similar or identical analysis in the other language. For example, if you learned to program R from that text, you will be able to learn Python rapidly using this book. And if you master analyses in Python here, you will be able easily to perform most of the same analyses in R using that text. You will already be familiar with most of the theoretical sections and datasets and can focus on the language. At the same time, each book has a few topics that are unique to its language and not covered in the other. This Python text has somewhat more emphasis on programming and writing custom functions. The R text includes methods for choice-based conjoint analysis (discrete choice models), market basket analysis with association rules, and methods to model behavior sequences such as weblogs. Those differences reflect the general situation in Python to have somewhat fewer yet often more stable and higher performance tools, contrasting an emphasis in R on a vast ecosystem of tools. In short, there are great reasons to learn both Python and R! The paired texts have been designed to make that easier. Acknowledgements We want to give special thanks here to people who made this book possible. First are all the students from our tutorials and classes over the years. They provided valuable feedback, and we hope their experiences will benefit you. Jason’s and Chris’s colleagues in the research community at Google provided extensive feedback on portions of the book. We thank the following Googlers: Javier Bargas, Mario Callegaro, Xu Gao, Rohan Gifford, Michael Gilbert, Xiaoyu He, Tim Hesterberg, Shankar Kumar, Kishan Panchal, Katrina Panovich, Michael Quinn, David Remus, Marta Rey-Babarro, Dan Russell, Rory Sayres, Angela Schörgendorfer, Micha Segeritz, Bob Silverstein, Matt Small, Gill Ward, John Webb, Rui Zhong, and Yori Zwols. Their encouragement and reviews have greatly improved the book. In the broader community, we had valuable feedback from Lynd Bacon, Marianna Dizik, Dennis Fok, Norman Lemke, Paul Litvak, Kerry Rodden, Steven Scott, and Randy Zwitch. The staff and editors at Springer helped us smooth the process, especially Senior Editor Lorraine Klimowich. Much of this book was written in public and university libraries, and we thank them for their hospitality alongside their unsurpassed literary resources. Portions of the book were written during pleasant days at the New Orleans Public Library, New York Public Library, Christoph Keller, Jr. Library at the General Theological Seminary in New York, University of California San Diego Giesel Library, University of Washington Suzzallo and Allen Libraries, Sunnyvale Public Library, and the Tokyo Metropolitan Central Library. Most importantly, we thank you, the reader. We are glad you have decided to investigate Python, and we hope to repay your effort. Let us start! Nashville, TN, USA Jason S. Schwarz Seattle, WA, USA Chris Chapman Philadelphia, PA, USA Elea McDonnell Feit August 2020
Авторы: Jason S. Schwarz, Chris Chapman, Elea McDonnell Feit (2020)
We are here to help you learn Python for marketing research and analytics. Python is a great choice for marketing analysts. It offers advanced capabilities for fitting statistical models. It is extensible and is able to process data from many different systems, in a variety of forms, for both small and large datasets. The Python ecosystem includes a vast range of established and emerging statistical methods as well as visualization techniques. Yet its use in marketing lags other fields such as econometrics, bioinformatics, and computer science. With your help, we hope to change that! This book is designed for two audiences: practicing marketing researchers and analysts who want to learn Python and students or researchers from other fields who want to review selected marketing topics in a Python context. What are the prerequisites? Simply that you are interested in Python for marketing, are conceptually familiar with basic statistical models such as linear regression, and are willing to engage in hands-on learning. This book will be particularly helpful to analysts who have some degree of programming experience and wish to learn Python. In Chap. 1, we describe additional reasons to use Python (and a few reasons perhaps not to use Python). The hands-on part is important. We teach concepts gradually in a sequence across the first seven chapters and ask you to type our examples as you work; this book is not a cookbook style reference. We spend some time (as little as possible) in Part I on the basics of the Python language and then turn in Part II to applied, real-world marketing analytics problems. Part III presents a few advanced marketing topics. Every chapter shows off the power of Python, and we hope each one will teach you something new and interesting. Specific features of this book are: • It is organized around marketing research tasks. Instead of generic examples, we put methods into the context of marketing questions. • We presume only basic statistics knowledge and use a minimum of mathematics. This book is designed to be approachable for practitioners and does not dwell on equations or mathematical details of statistical models (although we give references to those texts). • This is a didactic book that explains statistical concepts and the Python code. We want you to understand what we are doing and learn how to avoid common problems in both statistics and Python. We intend the book to be readable and to fulfill a different need than references and cookbooks available elsewhere. • The applied chapters demonstrate progressive model building. We do not present “the answer” but instead show how an analyst might realistically conduct analyses in successive steps where multiple models are compared for statistical strength and practical utility. • The chapters include visualization as a part of core analyses. We do not regard visualization as a standalone topic; rather, we believe it is an integral part of data exploration and model building. • Most of the analyses use simulated data, which provides practice in the Python language along with additional insight into the structure of marketing data. If you are inclined, you can change the data simulation and see how the statistical models are affected. • Where appropriate, we call out more advanced material on programming or models so that you may either skip it or read it, as you find appropriate. These sections are indicated by ∗ in their titles (such as This is an advanced section∗). What do we not cover? For one, this book teaches Python for marketing and does not teach marketing research in itself. We discuss many marketing topics but omit others that would simply repeat the analytic methods in Python. As noted above, we approach statistical models from a conceptual point of view and skip the mathematics. A few specialized topics have been omitted due to complexity and space; these include customer lifetime value models and econometric time series models. v vi Preface Overall, we believe the analyses here represent a great sample of marketing research and analytics practice. If you learn to perform these, you will be well equipped to apply Python in many areas of marketing. For another, this book teaches Python for marketing and not all of the complexity or subtlety of the Python programming language or of programming generally. We present the basics of Python programming that are needed to successfully analyze marketing data, but there is much more to Python than we present here. Companion Book: R for Marketing Research and Analytics This book is closely related to R for Marketing Research and Analytics (Chapman and Feit 2019) and shares many datasets and sections of didactic explanation of methods with that R text. In some ways, this Python book and the R text are mutual “translations” of one another from R and Python, respectively. This was a deliberate choice that we hope will make it easy for readers to move between Python and R. If you understand a method in one language, the companion book will demonstrate how to perform a similar or identical analysis in the other language. For example, if you learned to program R from that text, you will be able to learn Python rapidly using this book. And if you master analyses in Python here, you will be able easily to perform most of the same analyses in R using that text. You will already be familiar with most of the theoretical sections and datasets and can focus on the language. At the same time, each book has a few topics that are unique to its language and not covered in the other. This Python text has somewhat more emphasis on programming and writing custom functions. The R text includes methods for choice-based conjoint analysis (discrete choice models), market basket analysis with association rules, and methods to model behavior sequences such as weblogs. Those differences reflect the general situation in Python to have somewhat fewer yet often more stable and higher performance tools, contrasting an emphasis in R on a vast ecosystem of tools. In short, there are great reasons to learn both Python and R! The paired texts have been designed to make that easier. Acknowledgements We want to give special thanks here to people who made this book possible. First are all the students from our tutorials and classes over the years. They provided valuable feedback, and we hope their experiences will benefit you. Jason’s and Chris’s colleagues in the research community at Google provided extensive feedback on portions of the book. We thank the following Googlers: Javier Bargas, Mario Callegaro, Xu Gao, Rohan Gifford, Michael Gilbert, Xiaoyu He, Tim Hesterberg, Shankar Kumar, Kishan Panchal, Katrina Panovich, Michael Quinn, David Remus, Marta Rey-Babarro, Dan Russell, Rory Sayres, Angela Schörgendorfer, Micha Segeritz, Bob Silverstein, Matt Small, Gill Ward, John Webb, Rui Zhong, and Yori Zwols. Their encouragement and reviews have greatly improved the book. In the broader community, we had valuable feedback from Lynd Bacon, Marianna Dizik, Dennis Fok, Norman Lemke, Paul Litvak, Kerry Rodden, Steven Scott, and Randy Zwitch. The staff and editors at Springer helped us smooth the process, especially Senior Editor Lorraine Klimowich. Much of this book was written in public and university libraries, and we thank them for their hospitality alongside their unsurpassed literary resources. Portions of the book were written during pleasant days at the New Orleans Public Library, New York Public Library, Christoph Keller, Jr. Library at the General Theological Seminary in New York, University of California San Diego Giesel Library, University of Washington Suzzallo and Allen Libraries, Sunnyvale Public Library, and the Tokyo Metropolitan Central Library. Most importantly, we thank you, the reader. We are glad you have decided to investigate Python, and we hope to repay your effort. Let us start! Nashville, TN, USA Jason S. Schwarz Seattle, WA, USA Chris Chapman Philadelphia, PA, USA Elea McDonnell Feit August 2020
Скрытое содержимое могут видеть только пользователи групп(ы): Premium, Местный, Свои