Michael Siers

Book Review - Text Mining with R

Text Mining with R is very similar to the book R for Data Science (of which I'm a big fan). Whereas the latter is written by the creators of the tidyverse package, the former is written by the creators of the tidytext package.

We thus define the tidy text format as being a table with one token per row. - Text Mining with R, Julia Silge & David Robinson

The book starts off strong with captivating and easy-to-follow examples of creating tidy text dataframes from unstructured text documents, sentiment analysis, tf-idf, & n-grams. The chapters which follow on transformation to-and-from the tidy text format were not as exciting but provide a solid reference for projects. The book concludes with 3 case studies which give the reader an idea of what might constitute an end-to-end text analysis project.

I bought this book amongst several others on the subject of R. I love books with beautiful coloured visualisations and highlighted syntax such as my copies of R for Data Science and Forecasting: Principles and Practice. Unfortunately my copy of Text Mining with R was in black and white which made some of the visualisations hard to read. I gladly would have paid $20-$30 more just to have coloured print.

I'd recommend this book to data scientist who are new to text analysis, especially if they are familiar with the tidyverse. I have already begun sharing my newly gained knowledge at work on the matter and have a project in the pipeline.

My Ratings

Value for Money: ★★★★☆
Usefulness for Work: ★★★★☆
Expertise of the Authors: ★★★★★
Difficulty to Understand: ★★★☆☆
Printed Aesthetics: ★★☆☆☆

Average Score: 3.6 / 5.0