I decided to read “Social Media Mining with R”, because I am interested in the subject. Also the book is written by two authors with impressive careers in this field. It is unusual for a book this short (120+ pages) to be written by two experts. The most likely explanation is of course that they were too busy and wanted to share the workload. Both authors have PhDs, and as I said earlier impressive careers which also included DARPA projects. Therefore I assumed that the material would be spectacular of incredibly advanced technical level. Unfortunately I couldn’t be more wrong. The emphasis in this book turned out to be more on social science than on technical matters such as machine learning algorithms. Some chapters do not contain any code or anything of technical relevance. Therefore I feel it’s necessary to give a chapter-by-chapter overview to make it easy to skip chapters:
- The first chapter is introductory, but I would barely give it a passing grade, only because of some interesting factoids that are given about the evolution of social media. This is one of those chapters that I would skip.
- In the second chapter we finally get to see some code. It’s beginner level R code, which is fine by me, because I don’t know R that well. We also get instructions on installing and setting up a R development environment. If you are an experienced R coder I think you can safely skip this chapter.
- The third chapter is about Twitter and obtaining Twitter data. Some of the examples here are pretty good. But again the technical details do not get enough attention in my opinion.
- The fourth chapter is also theoretical without any code. Some of it seems relevant, but on the other hand it doesn’t seem that advanced or difficult.
- The fifth chapter explains a bit about Naive Bayes. The model the authors give on emotions and moods, however, seems like the sort of thing any average reader can come up with on their own. I mean we all have experience with moods and emotions in our daily life. Not so much with machine learning algorithms. Don’t get me wrong I am not saying that social science is a trivial pursuit. Just that the social science material presented in this book is not exactly rocket science.
- In the sixth chapter several case studies are presented with actual (finally) code. Most of the examples are US oriented, so it is assumed you know a bit about US politics and economics. The tutorials in this chapter seem to be of higher level. This lead me to believe that we were picking up the pace and I was looking forward to the next chapter. But there wasn’t a next chapter, if you don’t count the appendix that was it.
After reading this book I am left a bit disappointed, obviously because I had high (unrealistic) expectation. As I am personally not that interested in social science, I would have preferred that part to have been limited to an appendix and footnotes and I would have loved to see more code and technical background.