Statistical Inference via Data Science

Statistical Inference via Data Science: A ModernDive into R and the Tidyverse, Second Edition offers a comprehensive guide to learning statistical inference with data science tools widely used in industry, academia, and government. The first part of this book introduces the tidyverse suite of R packages, including ggplot2 for data visualization and dplyr for data wrangling. The second part introduces data modeling via simple and multiple linear regression. The third part presents statistical inference using simulation-based methods within a general framework implemented in R via the infer package, a suitable complement to the tidyverse. By working with these methods, readers can implement effective exploratory data analyses, conduct statistical modeling with data, and carry out statistical inference via confidence intervals and hypothesis testing. All of these tasks are performed by strongly emphasizing data visualization.

Key Features in the Second Edition:

• Minimal Prerequisites: No prior calculus or coding experience is needed, making the content accessible to a wide audience.
• Real-World Data: Learn with real-world datasets, including all domestic flights leaving New York City in 2023, the Gapminder project, FiveThirtyEight.com data, and new datasets on health, global development, music, coffee quality, and geyser eruptions.
• Simulation-Based Inference: Statistical inference through simulation-based methods.
• Expanded Theoretical Discussions: Includes deeper coverage of theory-based approaches, their connection with simulation-based approaches, and a presentation of intuitive and formal aspects of these methods.
• Enhanced Use of the infer Package: Leverages the infer package for “tidy” and transparent statistical inference, enabling readers to construct confidence intervals and conduct hypothesis tests through multiple linear regression and beyond.
• Dynamic Online Resources: All code and output are embedded in the text, with additional interactive exercises, discussions, and solutions available online.
• Broadened Applications: Suitable for undergraduate and graduate courses, including statistics, data science, and courses emphasizing reproducible research.

The first edition of the book has been used in so many different ways--for courses in statistical inference, statistical programming, business analytics, and data science for social policy, and by professionals in many other means. Ideal for those new to statistics or looking to deepen their knowledge, this edition provides a clear entry point into data science and modern statistical methods.

Chester Ismay is Vice President of Data and Automation at MATE Seminars and is a freelance data science consultant and instructor. He also teaches in the Center for Executive and Professional Education at Portland State University. He completed his PhD in statistics from Arizona State University in 2013. He has previously worked in various roles, including as an actuary at Scottsdale Insurance Company (now Nationwide E&S/Specialty) and at Ripon College, Reed College, and Pacific University. He has experience working in online education and was previously a Data Science Evangelist at DataRobot, where he led data science, machine learning, and data engineering in-person and virtual workshops for DataRobot University. In addition to his work for *ModernDive*, he contributed as the initial developer of the `infer` R package and is the author and maintainer of the `thesisdown` R package.

Albert Y. Kim is an Associate Professor of Statistical & Data Sciences at Smith College in Northampton, MA, USA. He completed his PhD in statistics at the University of Washington in 2011. Previously he worked in the Search Ads Metrics Team at Google Inc.\ as well as at Reed, Middlebury, and Amherst Colleges. In addition to his work for *ModernDive*, he is a co-author of the `resampledata` and `SpatialEpi` R packages. Both Dr. Kim and Dr. Ismay, along with Jennifer Chunn, are co-authors of the `fivethirtyeight` package of code and datasets published by the data journalism website FiveThirtyEight.com.

Arturo Valdivia is a Senior Lecturer in the Department of Statistics at Indiana University, Bloomington. He earned his PhD in Statistics from Arizona State University in 2013. His research interests focus on statistical education, exploring innovative approaches to help students grasp complex ideas with clarity. Over his career, he has taught a wide range of statistics courses, from introductory to advanced levels, to more than 1,800 undergraduate students and over 900 graduate students pursuing master's and Ph.D. programs in statistics, data science, and other disciplines. In recognition of his teaching excellence, he received Indiana University’s Trustees Teaching Award in 2023.