--- title: "Rmarkdown intro" author: "The ASTA Team" output: html_document: fig_height: 3 fig_width: 5 theme: cerulean highlight: tango pdf_document: fig_height: 3 fig_width: 5 --- # Introduction Rmarkdown First and foremost: Forget about everything above this line for now!!! This is a brief introduction to how Rmarkdown works (for more details on using R Markdown see ) using a few simple examples. Rmarkdown files consists of: - A header: title, author, output format etc. - Simple text like this (LateX users can exploit common LateX formulas like $\sqrt{9} = 3$ etc.). + Headings are created by hashtags; a single hashtag is the largest level. - Code chunks where we execute R commands. Code chunks begins with the line ```{r} and ends with the line ```. Everything in between are R calculations. - When you are done working on your project, you `knit` (the knit button is blue and is located at the top) the entire Rmarkdown file to either a pdf (requires that latex is installed on your computer) or a html file. A first peak at R chunks: ```{r} 1+1 ``` The above chunk can be executed by hitting the green "play" button to the right of the chunk or simply placing the cursor at line with `1+1` and hit Ctrl+Enter (on Windows). The result is shown in the "Console". ### Exercise: - Knit this Rmarkdown by clicking the "knit" button. This will run all R chunks consecutive (one after one). - Insert a new chunk either by using Ctrl+Alt+I or by clicking the green Insert button at the top: REMOVE THIS LINE AND INSERT A CHUNK - Once the chunk is inserted, calculate 1+2+3+4+5 Additionally Shortcuts can be found in the Help menu. # R ## R as a Calculator - **R** can be used like a simple calculator: ```{r} 3-5 3*9 6/5 sqrt(9) pi 3^2 log(100) ``` We can also include inline R code using backticks like this `r 1+4` and this `r sqrt(5)+3*7`. ## Variable Assignment - Assigning values to variables can be done in several ways: ```{r} a <- 99 b = 1 1.45 -> c ``` - If you hit the green button in the chunk above, you will see that the variables a,b and c are created in the environment pane (to the right). We can now work with these variables in subsequent chunks: ```{r} a+b+c ``` - Some prefer to use `=` for assignment, but the traditional way is by using `<-` which most users use. You will most likely never use `->`. - There are some restriction when assigning. For example the name of a variable can not start with a number. For longer names it is good practice to separate words with an underscore or using capital letters like: `mean_of_men_height` or `meanOfMenHeight`. However, such a long name for a mean value is probably overkill. + You may benefit from reading a short guide on code style like http://r-pkgs.had.co.nz/style.html + Meaningful names make your code easier for others to read, but also for yourself if you have to read your code again next month. - Make sure not to use "old" variable names (if a variable name is reused for a different value in the same document be cautious!) - To print the value of a variable, just write the name followed by "Enter" if in a console or "Ctrl+Enter" if in the editor: ```{r} a b c # This will print c in the final output ``` - In R (chunks), hashtags `#` allow comments and can be useful in reminding oneself of specific **R** code (i.e. what the code does). - Finally; we can also use variables in inline calculations like `r a+b-50`. ### Exercise: - Create the variables $x$ and $y$ and assign the value 23 to $x$ and 11 to $y$ and try to add these together by writing $x+y$. ```{r} # x = ... # y = ... # x + y ``` - Take the square root of y ```{r} # sqrt(...) ``` - Divide x by y ```{r} # ... / ... ``` # Data Structures - Vectors ## Types of Vectors The basic data structure in R is the vector which can be created using the `c()` function (for concatenate). ```{r} my_vector = c(1.3, 5, 3.5, 7) my_vector ``` ## Subsetting (indexing) Vectors - We can access the elements of a vector using the `[` function: ```{r} v = c(2.1, 4.2, 3.3, 5.4, 7) v[3] ``` - We can do even more advanced indexing using + Positive integers: Extracting specific elements + Negative integers: Exclude specific elements ```{r} v[c(1,3,5)] # Positive v[-c(1,3,5)] # Negative ``` ## Assignment - We can change elements in a vector using subsetting: ```{r} v2 = c(1, 2, 3) v2[1] = 0 v2 ``` - Another example: ```{r} v3 = c(3,5,7,9,11) v3[c(2,3)] = 0 v3 ``` ## Math Operations on Vectors R uses vectorized operations (addition, multiplication and logarithms of vectors etc.): ```{r} sum(v2) # Sum of all elements in v2 sqrt(v2) # Square root of v2 v2^2 # Squaring all elements in v2 v2 + v2 # Add v2 to it self ``` ### Exercise - Use R to calculate the sum of the numbers $1,2,3,4,5,6,7,8,9,10$ - Extract the second and third element of `v4 = c(4,3,9,1,0)` (you need to create a new chunk below). # Reading Data We shall now consider a specific dataset called `BrainSize`. We load it into R by the command ```{r} BrainSize = read.delim("https://asta.math.aau.dk/datasets?file=BrainSize.txt") ``` The `BrainSize` dataset is now of a form called a data frame in R; you can think of it as an excel sheet with variables as columns. ## Dataframes To get an overview (the first six rows) of a data frame you can use the `head()` function: ```{r} head(BrainSize) ``` We see that the `BrainSize` dataset consists of 7 columns: - `Gender` - `FSIQ` - `VIQ` - `PIQ` - `Weight` - `Height` - `MRI_Count` - `HeightIntervals`: The Height variable divided into 5 intervals We can extract columns (vectors) from a data frame with the `$` operator: ```{r} height = BrainSize$Height # Extracting the 'Height' column from data frame 'BrainSize' and naming it 'height' mean(height) ``` # Using Add-on Packages (Mosaic) R comes with a rich set of pre-installed functions like `mean`, `sum`, `plot` and many more. We can also install new "packages" with additional functionalities. Throughout this course, we shall rely heavily on the package `mosaic`. It can be installed be typing `install.packages("mosaic")` - you only have to do this the very first time. Whenever we want to use a function within the packages we need to load it. We only do this one time in the Rmarkdown document; and usually at the very beginning of the document. The command is `library(mosaic)`. In the rest of this tutorial, we shall develop an understanding of the usage of the `mosaic` package. You can look up important `mosaic` functions in the cheat-sheet or refer to this document. Many functions from `mosaic` have the form `goal(y ~ x | z, data = mydata, ...)`. For plots: - `y`: is the y-axis variable - `x`: is the x-axis variable - `z`: conditioning variable (separate panels) For other things: `y ~ x | z` can usually be read `y` is modeled by (or depends on) `x` differently for each `z`. ## Tabulate Recall the `BrainSize` data which we have already loaded into R. We can use the `tally` function from the `mosaic` package to summarize the Gender variable: ```{r, message=FALSE} # The option "message = FALSE" will prevent R from printing information about the package. library(mosaic) # The functionalities in the mosaic package are now available tally( ~ Gender, data = BrainSize) ``` What we see is, that in this data set $20$ observations are Females and $18$ observations are Males. We can also make a cross tabulation of `Gender` and `HeightIntervals` (remember that HeightIntervals is a categorical variable with five levels): ```{r} tally( ~ Gender + HeightIntervals, data = BrainSize) ``` With this command we simply "model" (count) `Gender` and `HeightIntervals` together. We see, that males are in general higher than females (for this data set). - To swap rows and columns swap the variables. - To get the relative frequencies (of the total observations) add `format = "percent"`: ```{r} tally( ~ HeightIntervals + Gender, data = BrainSize, format = "percent") ``` There is also an option to add totals in the "margins" of the table: ```{r} tally( ~ HeightIntervals + Gender, data = BrainSize, format = "percent", margins = TRUE) ``` To get relative frequencies for each gender (across columns) specify that you want `HeightIntervals` "modeled" (counted) by `Gender`: ```{r} tally(HeightIntervals ~ Gender, data = BrainSize, format = "percent") ``` ## Numerical summaries We can also summarize the `Height` by `Gender` numerically with the `favstats` function: ```{r} favstats(Height ~ Gender, data = BrainSize) ``` As we have already guessed, the mean height of men is larger than the mean of women height. We can also use the `mean` function to extract the means directly: ```{r} mean(Height ~ Gender, data = BrainSize) ``` ### Exercises Consider the BrainSize data. - What is the mean value of the Weight variable? - What is the mean value of the Weight variable for each group of the Gender variable? - How many females weigh $146$? ## Visualizing Data ### Boxplots ```{r} gf_boxplot(Height ~ Gender, data = BrainSize) ``` ```{r} gf_boxplot(Weight ~ HeightIntervals | Gender, data = BrainSize) ``` ### Exercises Consider the BrainSize data. - Use the function `gf_point` from `mosaic` to plot `Weight` against `Height` with a different color for each `Gender` (fill in the missing code). ```{r} # gf_point( ... ~ ..., col = ~Gender, data = BrainSize) ``` - Do the same, but with separate plots for males and females (fill in the missing code). ```{r} # gf_point( ... ~ ... | ..., data = BrainSize) ``` - Do you see a different picture for males and females?