Using R for SEO, What to expect?

The power of R. What's different about it?

if we were to talk about only one benefit of R, it's a High-level programming language that mainly focuses on data analysis. Meaning it's "specialized". With one or a few lines of code, you can do a lot. Let me give you an example:


This line of code, executed inside RStudio, will

  • prompt a select file menu for you to select a CSV (file.choose)

  • It will import data inside R (read.csv)

  • Display it (View)

Let's do it with a website links file

internal hyperlinks

This is how you open and browse a file with 2.6 Million rows effortlessly. Noticed the small search icon on the top right? Yes, you can search within it quite easily too.

search for dead links using http code

Want to count HTTP code? We will do this in two steps, first we load the same CSV file and save it as a variable. Nearly the same as before:

internal_linking = read.csv(file.choose())

Then we are going to display the count of Status column values


You can recognize the View function from before. the table function just count values. the $ is a shortcut to access column values

It displays:

count of http code

This is 30 secondes job. The most time-consuming part was to find the file on the hard disk.

Of course, these are just some silly examples. There is countless way to do this thing (third party app, terminal, excel pivot and of course python but it gives a nice intro of possibilities and how simple that is.

'There is a package for that'

The real power of R relies on R packages. What's a package you may ask? It's an on-demand library of functions you can load to help you in specialized tasks. Again let's take some examples.


It's one of the most famous packages. it can be used to build advanced charts and plots. To use it, you just have to install it once like this


to load it


and after that, you can now use it

aes(x = Status, fill = Status) +
geom_bar() +
scale_fill_hue() +

Because we only want to see the problematic http codes, we are going to filter

internal_linking_filtered <- filter(internal_linking, !(Status %in% c("200 no error", "Not checked","999 LinkedIn blocking automated testing")))
aes(x = Status, fill = Status) +
geom_bar() +
scale_fill_hue() +

Let's not go into details for now, but believe it or not, I'm not capable of writing this code, I just googled: "Bar charts chart ggplot" , "flip axis ggplot", ... shamelessly copy-paste the codes.

gggplot2 is powerful, it can make basically every chart you can think of

A few examples of plots done using ggplot2

To see more examples:

Let's look at another package


Lubridate will help to deal with our timestamp values. After the now-classic installing and loading


It can be used to guess and transform this Time.stampinto a real date format

internal_linking$real_date = dmy_hms(internal_linking$Time.stamp)

Values have been transformed into a true Date format.

before and after using Lubridate function

No more "at" in the middle or "am/pm". It's now easier to read and to sort. The function guessed successfully that the "at" was useless.

Now that those are real dates and no longer character string, we can plot them using ggplot

ggplot(internal_linking) +
aes(x = real_date) +
geom_histogram() +

the Lubridate package can also help with duration, time zone, intervals, ... Have a look at the cheatsheets. It is a bit complex to get into but so much less than trying to do it yourself. I've lost literally days of my working life, trying to do this kind of stuff badly in Excel/Google Sheet.


One last example for the road. 'Want to extract links domains? You can sure use regex, or even try to split the string using "/" as a separator... OR you can use the more reliable urltools package which as a dedicated domain() function to do exactly that.

# Installing and Loading Package
# extract domain and feed it to a new data column called 'domain'
internal_linking$domain <- domain(internal_linking$URL)

Let's check out the values, nearly the same code as before:

top domains

Where to find packages?

Good question! All the previous packages have been downloaded from CRAN. It's a repository that contains thousands of packages. Github is also a great source. There are so many that, the problem is often to find the right one. The way to go is usually to ask around using:

The community is smaller than other programming languages but people are more willing to help, it compensates.

The confusing things about R

The name

Oh you do 'R programming', that's cool. Is it like Air Guitar? You do fake programming? - An anonymous member of my family

"R" is a weird name, especially in this covid time, and it's not the most Google-friendly name either. So here are few links to help find R resources.

the <-

If you've seen some R' code before and you might have been surprised to see this "<-" being used. it's just a legacy thing, historically R differentiate "assignation" and "comparison", example:

assignation - If you want to set the value of X to 3.

x <- 3

comparison - Is X equal to 3?

if (x = 3) {
return "the value of X is 3"

If you want to keep this little tradition alive you can use <- but it is really up to you. Perfectly fine to use =

x = 3
x <- 3

(these are actually little differences between the two but if you are new to R it's not important at all.)

The %>%

The %>% operator allows operations to be carried out successively. Meaning the results of the previous command are the entries for the next one. Like the > ( “pipe”) command line for terminal if you came across it.

Always better with an example, let's take the first line of code of this page


It's 3 functions being used on top of each other. The readability is decent. I wouldn't recommend adding a fourth. the %>% operator fixes this soon-to-be problem.

# equivalent to the previous instruction
file.choose() %>% read.csv() %>% View()
# again equivalent
file.choose() %>%
read.csv() %>%

As you can see, fairly easy to read. This operator is so practical that it's now used by a majority of R practicers.

R' can be slow

I'm going to quote Hadley Wickham on this one:

R is not a fast language. This is not an accident. R was purposely designed to make data analysis and statistics easier for you to do. It was not designed to make life easier for your computer. While R is slow compared to other programming languages, for most purposes, it’s fast enough.

My personal experience is every time my code was slow, it was because I was doing something the code shouldn't have done, like recomputing all data to output a new record. if you are interested in performance issue, here is a website dedicated to Put R on Prod