Telling Stories with Data: With Application in R. By RohanAlexander. CRC Press. 2023. 622 pages. AU$129.60 (hardback). ISBN: 978‐1‐0321‐3477‐2.
Year: 2024
Type: book-review
Abstract: This book delves into various aspects of a data science workflow concretely illustrated with examples tailored for social science applications.Each example contains the full analysis code in the programming language R (R Core Team 2023)., primarily using the tidyverse (Wickham et al. 2019) paradigm, and instructions for obtaining the data where applicable.The book is not just about data analysis but places the human element in its core, with frequent reminders of reproducibility, ethical issues and mindset needed for tackling data projects.The book contains 17 chapters divided into six parts: Foundations, Communication, Acquisition, Preparation, Modelling and Applications.Each chapter starts with a list of reading materials or videos and R packages used in the chapter, and ends with a set of exercises and further reading materials.The book is available online at https:// tellingstorieswithdata.com/ and includes an additional part, Appendices, which is available online only.Foundations introduces data science workflow advocated by the author (Plan → Simulate → Acquire → Explore → Share) and integrates this workflow into examples throughout the entire book.A reproducible workflow is introduced, with the author recommending several tools and approaches, e.g.Quarto (Allaire et al. 2024), a particular file and folder structure with R projects, coding style in the tidyverse (Wickham et al. 2019) paradigm, and Git and GitHub for version control.Unfortunately, the screenshots showing these tools' usage are too small in print to follow along.Communication describes the aspects of writing research and static forms of data communication.In formulating the research question of interest, the author describes broadly two approaches: data-first and question-first.In writing research, the author advocates for writing frequently and completing the first draft as quick as possible.Each component of a paper (title, abstract, introduction, data, model, results and discussion) is briefly discussed.In static communication, the authors show different types of visualisation, with a special emphasis on maps.Code demonstrates the formatting of tables and easy construction of table with summary statistics and regression results.Acquisition discusses the imperfections of the measurement process, demonstrates obtaining Canadian census data and US census data via R packages, and provides steps to extract the American Community Survey via the IPUMS website (though the screenshot is unfortunately too small in print to see).It illustrates different sampling schemes with code and visualisation (systematic sampling, simple random sampling, stratified sampling and cluster sampling).In addition, it also demonstrates how to obtain data from the web, APIs and PDFs, including the use of Optical Character Recognition (OCR) to extract data from images.Preparation demonstrates how to clean, tidy, check, test, document, store and share data.Data fidelity is described in terms of validity, internal consistency and external consistency.Some tedious but important aspects of data cleaning are demonstrated with
DOI: 10.1111/anzs.12428
Author: Emi Tanaka
Institution: Australian National University
Language: English
Topic: Data Analysis with R
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4402999229