Time Series for Data Scientists: Data Management, Description, Modeling and Forecasting

by Juana Sanchez

Overview

Learn by doing with this user-friendly introduction to time series data analysis in R. This book explores the intricacies of managing and cleaning time series data of different sizes, scales and granularity, data preparation for analysis and visualization, and different approaches to classical and machine learning time series modeling and forecasting. A range of pedagogical features support students, including end-of-chapter exercises, problems, quizzes and case studies. The case studies are designed to stretch the learner, introducing larger data sets, enhanced data management skills, and R packages and functions appropriate for real-world data analysis. On top of providing commented R programs and data sets, the book's companion website offers extra case studies, lecture slides, videos and exercise solutions. Accessible to those with a basic background in statistics and probability, this is an ideal hands-on text for undergraduate and graduate students, as well as researchers in data-rich disciplines

Product Details

ISBN-13:	9781108837774
Publisher:	Cambridge University Press
Publication date:	05/11/2023
Pages:	550
Product dimensions:	6.89(w) x 9.84(h) x 1.02(d)

About the Author

Juana Sanchez is Senior Lecturer in Statistics at the University of California, Los Angeles. She is Editor of the Datasets and Stories section of the ASA's Journal of Statistics and Data Science Education and is the author of Probability for Data Scientists (2020).

Table of Contents

Part I. Descriptive Features of Time Series Data: 1. Introduction to time series data; 2. Smoothing and decomposing a time series; 3. Summary statistics of stationary time series; Part II. Univariate Models of Temporal Dependence: 4. The algebra of differencing and backshifting; 5. Stationary stochastic processes; 6. ARIMA(p,d,q)(P,D,Q)$_F$ modeling and forecasting; Part III. Multivariate Modeling and Forecasting: 7. Latent process models for time series; 8. Vector autoregression; 9. Classical regression with ARMA residuals; 10. Machine learning methods for time series; References; Index.

Editorial Reviews

‘This book provides an excellent introduction to time series modelling and forecasting which are increasingly important tools in the domain of official statistics. The clear descriptions and real-life examples provided in this text make it easy to digest for those not already familiar with the topic. In addition, the exercises allow readers to develop their understanding in more depth through hands-on applications of the methods to real data using open-source tools. The inclusion of modern topics such as machine learning and artificial intelligence are a valuable addition to make the text relevant and comprehensive.’ Steve Matthews, Statistics Canada

‘This book is a great introduction to the ideas and methods of time series data analysis. Chapter by chapter, it will show you its most valuable features, like the wealth of real examples as well as practical uses of R and graphical visualization. You will certainly enjoy this text, as it is suitable for a wide range of statistical courses.’ Vera Ioudina, Texas State University

‘Lots of good real world examples together with the use of R helps a lot as do the nice set of exercises. In time series, it is a tricky balance between overdoing theory or just hand waving and here the author does very well. This would make a lovely course text!’ Gareth Janacek, University of East Anglia

‘Time Series for Data Scientists’ develops your intuition before walking through classical and modern time series methods in easy-to-understand terms. With each algorithm Dr. Sanchez first helps you understand the motivation behind the approach; then walks you through the formulas step-by-step, outlining what we're doing and why; she also includes R code to help you apply the techniques learned to solve real-world business problems using real-world data sets; and takes the time to show you how to interpret the output, and discuss what to try next when an initial approach doesn't quite match the trends in the data. Whether you're an undergraduate or graduate student, are curious about time series methods, are looking for a self-paced book, or a reference guide, this is a must-have.’ Irina Kukuyeva, Fractional Chief Data Officer

‘A fine textbook for an introductory time series course aimed at undergraduates in Statistics or Data Science. The author did an excellent work in the choice of topics, covering from classical exploratory techniques to modern machine learning approaches, while keeping the level of the exposition accessible to readers with a modicum of mathematical background. To be recommended!’ Giovanni Petris, University of Arkansas

‘This book should be a serious contender if you are looking for an introductory text for an undergraduate course in time series. It is especially suited for a course populated with students having varying degrees of mathematical skill levels. Its conversational approach to introducing time series concepts and the use of insightful examples throughout the book makes it very accessible to students who are not highly trained in abstract mathematical reasoning. Nevertheless, it does not shy away from providing the theoretical underpinnings of various time series models but does so in a manner very accessible to students. The availability of R code throughout the book is an added plus. Even if I am teaching an upper-level graduate course in time series, I would use this book as a supplement simply because of the plethora of examples and data sources it provides.’ V. A. Samaranayake, Missouri University of Science and Technology

From the Publisher