Install R and RStudio Desktop

We will be using the programming language R and the software RStudioDesktop, which serves as an editor for your code but also comes with abunch of additional features.

1. Download and installR foryour operating system. On Windows, choose thebaseversion.
2. Download and installRStudioDesktop

For more information about R programming have a look at the freeonline bookHands-OnProgramming with R by Garrett Grolemund, which has a chapter withdetailedinstallationinstructions. For a quick video tour of the RStudio Desktop userinterface check outRStudio for the TotalBeginner.

Install packages

After opening RStudio, you will find your R console in the lower leftcorner. All code you enter in the console will be directly executed byR. Let’s start by installing some essential packages. Packages deliveradditional functionality that is not built into base R.

install.packages("tidyverse")install.packages("tidyplots")

library(tidyplots)study#>    treatment     group dose participant age    sex score#> 1          A   placebo high         p01  23 female     2#> 2          A   placebo high         p02  45   male     4#> 3          A   placebo high         p03  32 female     5#> 4          A   placebo high         p04  37   male     4#> 5          A   placebo high         p05  24 female     6#> 6          B   placebo  low         p06  23 female     9#> 7          B   placebo  low         p07  45   male     8#> 8          B   placebo  low         p08  32 female    12#> 9          B   placebo  low         p09  37   male    15#> 10         B   placebo  low         p10  24 female    16#> 11         C treatment high         p01  23 female    32#> 12         C treatment high         p02  45   male    35#> 13         C treatment high         p03  32 female    24#> 14         C treatment high         p04  37   male    45#> 15         C treatment high         p05  24 female    56#> 16         D treatment  low         p06  23 female    23#> 17         D treatment  low         p07  45   male    25#> 18         D treatment  low         p08  32 female    21#> 19         D treatment  low         p09  37   male    22#> 20         D treatment  low         p10  24 female    23

library(tidyplots)

study|>tidyplot(x=treatment, y=score)

Add

Next, let’s add some more elements to the plot. This is done by usinga family of functions that all start withadd_. Forexample, we can add the data points by adding one more line to the code.Note, that we need a|> at the end of each line, wherethe output should be piped into the next line. When you combine multiplelines like this, you have generated apipeline.

study|>tidyplot(x=treatment, y=score)|>add_data_points()

Of course, you do not have to stop here. There are manyadd_*() functions you can choose from. An overview of allfunction in the tidyplots package can be found in thePackageindex.

For now, let’s add some bars to the plot. As soon as you start typing“add” in RStudio you should see a little window next to your courserthat shows all available function that start with “add” and can thus beused to build up your plot. You can also manually trigger theauto-completion window by hitting thetab key.

In tidyplots, function names that start withadd_usually continue with the statistical entity to plot,e.g. mean,median,count, etc. Asa next piece, you decide which graphical representation to use,e.g. bar,dash,line etc. In ourexample we chooseadd_mean_bar() to show the mean value ofeach treatment group represented as a bar.

study|>tidyplot(x=treatment, y=score)|>add_data_points()|>add_mean_bar(alpha=0.4)

One thing to note here is that I addedalpha = 0.4 as aargument to theadd_mean_bar() function. This adds a littletransparency to the bars and results in a lighter blue color incomparison to the data points.

Some people might do not like bars so much. So let’s exchange thebar for adash. And while we are on it, let’sadd the standard error of the meansem, represented aserror bar.

study|>tidyplot(x=treatment, y=score)|>add_data_points()|>add_mean_dash()|>add_sem_errorbar()

I think by now you got the principle. You can just keep adding layersuntil your plot has all the elements you need.

But there is one more building block that we need to cover and thatis color. Color is a very powerful way to encode information in a plot.As colors can encodevariables in a similar way as axes, theargumentcolor needs to be to provided in the initial callof thetidyplot() function.

study|>tidyplot(x=group, y=score, color=dose)|>add_data_points()|>add_mean_dash()|>add_sem_errorbar()

As you can see,color acts as a way to group the data bya thirdvariable, thus complementing thex andy axis.

Although there are many moreadd_*() functionsavailable, I will stop here and leave you with thePackageindex and the article aboutVisualizingdata for further inspiration.

Remove

Besides adding plot elements, you might want to remove certain partsof the plot. This can be achieved with theremove_*()family of functions. For example, you might want to remove the colorlegend title, or in some rare cases even the entire y-axis.

study|>tidyplot(x=group, y=score, color=dose)|>add_data_points()|>add_mean_dash()|>add_sem_errorbar()|>remove_legend_title()|>remove_y_axis()

Moreremove_*() functions can be found in thePackageindex.

Adjust

After you have assembled your plot, you often want to tweak somedetails about how the plot or its components are displayed. For thistask, tidyplots provides a number ofadjust_*()functions.

Let’s start with this plot.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()

When preparing figures for a paper, you might want ensure, that allplots have a consistent size. The default in tidyplots is a width of 50mm and a height of 50 mm. Please note that these values refer to size oftheplot area, which is the area enclosed by the x and y-axis.Therefore labels, titles, and legends are not counting towards theplot area size.

This is perfect to achieve a consistent look, which is most easilydone by selecting a consistentheight across plots, whilethewidth can vary depending on the number of categories inthe x-axis.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points_beeswarm(shape=1)|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>adjust_size(width=20, height=20)

Another common adjustment is to change the titles of the plot, axes,or legend. For this we will use the functionadjust_title()and friends.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>adjust_title("This is my fantastic plot title")|>adjust_x_axis_title("Treatment group")|>adjust_y_axis_title("Disease score")|>adjust_legend_title("")|>adjust_caption("Here goes the caption")

Note that I removed the legend title by setting it to an empty stringadjust_legend_title(""). This is alternative toremove_legend_title(), however the result is not exactlythe same. I am sure you will figure out the difference.

Another common task is to adjust the colors in your plot. You can dothis using theadjust_colors() function.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>adjust_colors(new_colors=c("#644296","#F08533","#3B78B0","#D1352C"))

You can also use the color schemes, that are built into tidyplots. Tolearn more about these color schemes have a look at the articleColorschemes.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>adjust_colors(new_colors=colors_discrete_seaside)

Rename, reorder, sort, and reverse

A special group of adjust functions deals with thedatalabels in your plot. These function are special because they needto modify the underlying data of the plot. Moreover, they do not startwithadjust_ but withrename_,reorder_,sort_, andreverse_.

For example, to rename the data labels for thetreatmentvariable on the x-axis, you can do this.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>rename_x_axis_labels(new_names=c("A"="This","B"="is","C"="totally","D"="new"))

Note that we provide anamed character vector to make itclear which old label should be replace with which new label.

The remaining functions, starting withreorder_,sort_, andreverse_, do not change the name ofthe label but their order in the plot.

For example, you can bring the treatment “D” and “C” to thefront.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>reorder_x_axis_labels("D","C")

Sort the treatments by their score.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>sort_x_axis_labels()

Or simply reverse the order of the labels.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_mean_bar(alpha=0.4)|>add_sem_errorbar()|>reverse_x_axis_labels()

Of course, there are many moreadjust_ functions thatyou can find in thePackageindex.

Themes

Themes are a great way to modify the look an feel of your plotwithout changing the representation of the data. You can stay with thedefault tidyplots theme.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_sem_errorbar()|>add_mean_dash()|>theme_tidyplot()

Or try something more like ggplot2.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_sem_errorbar()|>add_mean_dash()|>theme_ggplot2()

Or something more minimal.

study|>tidyplot(x=treatment, y=score, color=treatment)|>add_data_points()|>add_sem_errorbar()|>add_mean_dash()|>theme_minimal_y()|>remove_x_axis_line()

Split

When you have a complex dataset, you might want split the plot intomultiple subplots. In tidyplots, this can be done with the functionsplit_plot().

Starting with thestudy dataset, you could plot thescore against the treatmentgroup and splitthis plot bydose into a high dose and a low dose plot.

study|>tidyplot(x=group, y=score, color=group)|>add_data_points()|>add_sem_errorbar()|>add_mean_dash()|>adjust_size(width=30, height=25)|>split_plot(by=dose)

Output

The classical way to output a plot is to write it to a PDF or PNGfile. This can be easily done by piping the plot into the functionsave_plot().

study|>tidyplot(x=group, y=score, color=group)|>add_data_points()|>add_sem_errorbar()|>add_mean_dash()|>save_plot("my_plot.pdf")

Conveniently,save_plot() also gives back the plot itreceived, allowing it to be used in the middle of a pipeline. Ifsave_plot() is a the end of pipeline, the plot will berendered on screen, providing a visual confirmation of what was saved tofile.

tidyplots documentation

• Packageindex
  Overview of all tidyplots functions

• Getstarted
  Getting started guide

• Visualizingdata
  Article with examples for common data visualizations

• Advancedplotting
  Article about advanced plotting techniques and workflows

• Colorschemes
  Article about the use of color schemes

Other resources

• Hands-OnProgramming with R
  Free online book by Garrett Grolemund

• R for Data Science
  Free online book by Hadley Wickham

• Fundamentals of DataVisualization
  Free online book by Claus O. Wilke