Writing a R book and self-publishing it in Amazon
My experience and some advices about the process
Many people, including my university colleagues and friends, have asked me about the process of writing a book and self publishing it in Amazon. You can find the details about the english version of the book here and here. Given so much interest, I’m going to report the whole process in this post.
First, motivation. Why did I write a book?
I am a university professor. Writing is a major part of my work and I really enjoy it. Think about it, it is a magical process. I press a specific and long combination of strokes in my keyboard and that translates into information distributed all over the world. This information helps people in a particular problem, just as I have been helped in the past by reading the work of other people.
A longer answer, I wrote the book because I simply felt like studying and learning more about R. Writing a book is an excellent opportunity for doing so as it forces you to think about a topic and communicate it clearly and objectively. After reading the book from Grangrud, I got some inspiration for writing my own book about R, reproducible research and Finance. My experience in the academia tells me that people often learn programming by themselves, without any consideration for the structure and re-usability of the code. You can see examples of it when you download code from other researchers. In most cases, it is a big mess and I often found it easier (and less risky) to rewrite it all from scratch. So, writing a book is a way to show to people how to use R for data analysis in finance and how to organize code that can be used later, saving lots of time of development. As I mentioned in a previous post, another argument for writing the book is that, even if I didn’t publish it, I would still have very good class notes for my graduate students or perhaps a pdf to host in my website.
Off course you may wonder about the financial side. After all, I am a finance professor and finance is the science of money. While you can make money selling books, the financial incentives for writing a new technical book are quite low. In financial terms, writing a technical book is a project with negative expected Net Present Value (NPV). This means that, from the ex ante point of view of a potential writer, working in a technical book simply does not create financial value. There is a high production cost in terms of how much time you spend writing and formatting the book and a high uncertainty about the future royalties. If you look at it from the financial side, it is just a bad gamble. This is why you should only write a book about something you love and want to learn more about. Otherwise, it can become a big frustration.
The writing process
The first step in writing my book was to set out a summary, that is, the names of the chapters. The first part of it was pretty obvious, it covered the basics of using R: packages, objects, functions, loops and so on. The rest of the topics were all based on my experience. I asked the question What are the most important data tasks that a finance student should learn?. Once you have a TOC (table of contents) ready, it is all about filling up the book with content.
I really enjoyed the process of writing the book. Taking a long term project such as a book requires discipline. One strategy that really helped me was to work very early in the morning. As soon as the sun rised at 0630, I’m up and in my way to the university. From 07:00 am to 0900 am, nothing really happens at the university, giving me plenty of time to write peacefully. Some days I could write all morning, if I felt like it and had a free schedule. One tip for anyone that is thinking about writing a book (or article), only write when you fell like writing. It is really difficult to write well if your mind is not into it. Not having a deadline for this project meant that I could take my time in developing it.
As for software, the first version of the book was written in TexStudio with the content in
Sweave files (.Rnw), and a simple hack to call Sweave from R and compile the resulting tex file. I’m used to work with latex, so the choice for textudio was obvious. I like this latex editor as it integrates nicely with grammar checking tools and latex compilation. Why not
knitr? Well, at the time I was very comfortable with
Sweave. It offered everything I needed. I didn’t see a reason for change. Clearly, I was wrong. Once I started to investigate how to format better the code in the book, it became clear that I should be using
knitr, which has all these extra options that are not available in
Sweave. I switched it as soon as I realized how much I would gain there was.
After 6 months of work, I had a first readable version of the book. This is when I started to investigate how to turn a latex file into an ebook file. To my surprise, this is not easy! It could be the case that I had lots of code and figures from the compilation. I tried all the existing software and, to my frustration, it just didn’t work. When I didn’t got an error message that stopped the conversion, the resulting epub file look awful in the kindle reader. The warning and error messages from the latex2ebook compiler also didn’t help a lot.
By looking for solutions in the internet and with a bit of luck, I’ve found that Yihui Xie had just published his bookdown package. I tried out the first chapters and it worked perfectly! The great thing about
bookdown is that you can output to pdf, html or epub with the same files. I was very happy that, finally, I would be able to get all formats that I needed in a single platform. I’m not going to go into the details about how to use
bookdown. You can find the official tutorial here. It is worth to point out that the written content is just RMarkdown so, if you know latex, you’ll fell right at home. So, here I go again, switching a whole book from latex to Rmarkdown.
The only feature in
bookdown that didn’t work out of the box was the equation in the kindle format. The formulas just didn’t print well. The weird part is that they looked fine in any other ebook reader, but not in Kindle. Apparently, there is no solution for this problem besides using figure files for the equations. What I did to solve it was to use R switches within the book so that the epub compilation used a figure file created in codecogs and the latex/html compilation used the normal code for equations. See here an example of code for the first equation in chapter 9.
As you can see, I used a switch in the
knitr command as in
include=identical(knitr:::pandoc_to(), 'epub3') to run each code conditional to the type of output. As for inline equations, I just used straight text to indicate coefficients such as beta, alpha. I know I could have used the same strategy, but I didn’t fell the need since there were so few inline equations in the text.
After switching to bookdown, I added a couple more chapters in the following months. A comment here is that I tried writing in RStudio but it didn’t worked out for me. Perhaps it was due to the large files I was handling, but it felt really clunky, slow and not reactive. Also, the grammar check system in Rstudio needs a lot of work. I tried going back to Texstudio but it didn’t had support for markdown highlighting. I then switched software by doing the writing in notepad++ + markdown extension and only used RStudio to compile the book. If you haven’t tried notepad++, give it a go. I really like it and I find myself using it more and more.
Publishing in Amazon
Around october of 2016, the content of the book started to take a good shape and it was time to investigate where and how to publish it. I got in touch with a major local publisher here in Brazil and, after one month, they informed me that they were not interested in the book. While I could try other publishers, I really didn’t felt like going through another month of evaluation. I studied for while and decided to self-publish the book.
One positive aspect of self publishing is that big publishers downsized over the years and good professionals are now in the market for hire. This means that a large range of good editors and designers are available to the independent author. In my case, I fought the urge to do the cover myself and searched for a cover designer. After a lot of search, I hired Capista to do the book cover. I really liked his portfolio and, he offers a discount for self-publish authors, which is very nice of him.
I also hired professionals to check the text for mistakes and grammar. This is particular important as grammar mistakes are normal and expected in large documents. But, selecting a reviewer was an interactive process. I searched the web and hired tree different reviewers and offered them three different chapters of the book. Once I had all revisions, I selected the reviewer the offered the best result and sent him the rest of the chapters. You can find his site here. I’m glad I did this way, It would be a nightmare if I hired the wrong person to do this job.
The main advantage of self publication is that the royalties for the author are higher and the cost to publish are lower. This setup translates into a lower price of the book and more sells. Specific to Brazil, where the exchange rate of dollar to Real (the Brazilian currency) is very unfavourable, it makes sense to offer a lower priced book. This way, it is easier for students and instructor to acquire and use the book.
Since ebooks is a growing market, you can find lots of tutorials about this topic in the internet, like this one and this one. The problem here is that they are mostly about self-publishing text ebooks, as in fiction stories. The only article I’ve found about self-publishing technical books is this one, where the author describe their rather good experience in publishing a software-engineering textbook.
I checked Lulu, Smashwords and Amazon KDP. What made my choice towards Amazon was that it was a established platform, where almost everyone is registered. Buying an Ebook in Amazon is just one click away. Also, I’ve found that the KDP (self publish) program of Amazon is quite good. If the ebook sells for less than 9 dollars and is enrolled in the Amazon unlimited program, the author gets 70 percent of the sticker price in royalties, which is a nice percentage! A bit of information, the normal royalty rate for an author of a publishing company is around 5-10 percent of the book price. Just a quick comparison, for each ebook sold at 9 USD, the author gets 6.3 USD (0.70 times 9). In other scenario, assuming an author has a publisher that offered 10 percent royalty, you would get the same royalty for a a book priced at 63 USD (6.3/0.1). The big difference here is that a 9 USD ebook sells a lot more than a 63 USD ebook, resulting in higher royalties and higher impact. More people read your book and you make more money. My only disappointment with Amazon is that the printed books are shipped from the USA. A buyer of the printed book in Brazil will need to wait at least 10-15 days to get the book. But, overall, I find this to be a small cost to bear.
When the final version of the content of the book was ready, I started to organize everything I needed in order to publish the book. This is the list:
A ISBN number - A unique id that indexes your book, one for the ebook and another for the print version. In order to get an ISBN, you need to find the representation of ISBN in your country. Amazon also offers their own ISBN, but I don’t advise to use it. Having your own ISBN gives you more control. Just get one yourself. It is easy and cheap. In my case, I sent the request from the website by filling up forms in less than 10 minutes and got the number in five days.
Cover - The cover is processed as a figure file (.jpg). An important information here is that you need to define the size of you book and the number of pages at this stage. In my case, I used a 7x11 inches configuration for 200+ pages. It looks quite nice. Have a look:
Beautiful cover! Isn’t it? Once you have the figure file, you just add it to the book creator system in amazon. At this stage, you will also need a summary about the book and its author and a picture for the back cover of the printed version.
Registration in Amazon. The site asks you a lot of information. It took me at least half an hour to fill up the forms, which also includes tax details for us and non-us residents.
Description of the book. This is the text that goes in the webpage of the book in Amazon. In my case, I just used the same text as the back cover.
A book site. I wanted to distribute the code from the book and also exercises over the web. I used Google Sites, but I’m sure you have plenty of options here. On a side note, I’m also considering writing a CRAN package for distributing this content.
For the ebook, the next step is easy. Just upload the cover, the
.epub file from
bookdown, write the description and other details and you are good to go. With everything ready to go, you can check the ebook page by page with the online viewer. I didn’t have any problems at this stage. As for the print version, it is far more complicated. In my case, I used the pdf file to build the printed version. Once you upload the pdf, you can check the result in a online viewer. This is the final proof of the book and shows how it would look like when printed. There are lots of error checking from the system and, if you got an error, you cannot publish the book. I’ve spent a lot of time formatting everything so that the book comes out perfectly. The problem is that, each attempt to solve errors demanded processing time. At least 10 minutes to reach the whole cycle of compiling the pdf, uploading and checking the result. So, every time I tried to solve an error, I got the result 10 minutes later. I repeated this process over and over until I had a perfect book.
The book was launched in the sixth of February 2017. I advertised it in here, my blog, Facebook and email. I was very surprised with the strong reception. In the first and second day, it became the best seller book in the section of finance in the Brazilian Amazon. I am really happy for the success of the book. Right now, the sales have gone down, which is normal and expected, but I’m getting all sorts of good responses. What surprised me is that, even though the book was published primarily for Brazil, I’m receiving lots of replies (and sales) of people from other Portuguese speaking countries such as Cabo Verde and Portugal!
If I could go back in time and advice myself about the book, this is what I would have said. I hope these advices are helpful for others as well.
From the first version, make up your mind about book size and letter size. This decision will impact everything else. Problems are easier to detect and solve if you stick to a single format.
bookdownare your friends. I’m not aware of any other free platform that offers anything like it for developing a technical book.
When possible, use
tidy=TRUEin the chunks of code. This simple command forces the code within the boundaries of the pdf. Sometimes it doesn’t look good so you always need to mannualy check the chunks with lots of code.
Avoid using function
strin the content. For some reason, it does not respect the boundaries of the page. There is a solution but it would not look good in the book.
Use a dark theme for writing. Your eyes will hurt less and you will be able to write more.
Talk to people that understands the material you are writing about. Listen to their advices and make changes when necessary. Don’t over commit to sections that you are not sure will be included.
For all figure files in your book, always get the highest possible resolution (at least 300 dpi). Amazon will not let you publish figures with low resolution.
Don’t even try to do yourself the book cover or grammar check. Hire someone to do it for you. Your future self will thank you later!
I really enjoyed the experience of self-publishing a book and I strongly advise for anyone to give it a try. A cautionary note, I never worked with a established publisher and, therefore, I cannot talk about how it is or compare it to self-publish. If someone has had that experience, fell free to use the comments to discuss it.
The impact of a technical book is uncertain. Only time will tell if the book is successful or not. I have high hopes though. The feedback from the community has been strongly positive. I also have been working on extra material such as exercises and slides that are going to be distributed for free. This should motivate professors to use the book in class and increase distribution.
I hope this post was helpful and enlightening for potential authors. Comments or suggestions, please use the comment section or drop me an email at firstname.lastname@example.org.