In jorvlan/openvis: Create visualizations based on best open science practices

Background

This tutorial is partly inspired by work from Allen et al. (2019) and was initially created to contribute to a GitHub repository called 'open-visualizations'. The idea behind the 'open-visualizations' repository stems from the fact that (open) science - in general - lacks 'fully' transparent and robust visualizations, i.e., figures have always some form of 'hidden-data'. To overcome this issue, I started experimenting with some code in Python and R and after posting a Tweet, in which I stated that I was working on 'open-visualizations', Neuroscientist Micah Allen replied and advised me to check out his work on Rainclouds. After performing two of their tutorials, in R and Python respectively, I thought of a way to combine the raincloud approach with my own work performed in Python. Shortly after posting another Tweet, which included some figures that I produced in R, I received a huge amount of encouraging feedback which led me to writing this markdown document. Next to this markdown document, I have created two other tutorials in Python. Both tutorials are available in my repository open-visualizations.

If you have any questions, suggestions for improvement or identify bugs, please open an issue in the GitHub repository open-visualizations.

If you use my repository for your research, please reference it.

knitr::opts_chunk$set(fig.width=6, fig.height=5, fig.path='figs_repmes/',
                      echo=TRUE, warning=FALSE, message=FALSE)

R-version check

R.version$version.string

Package dependencies

Make sure you have the packages that are needed for this tutorial.

• Install plyr before dplyr. If you need functions from both plyr and dplyr, please load plyr first, then dplyr, otherwise error messages might occur (source: R console). Rmisc also depends on this package.
• Install lattice since Rmisc dependends on this package.
• Install ggplot2, dplyr, and readr.
• Install rmarkdown to convert this .Rmd template into a variety of formats including HTML, MS Word, PDF, and Beamer (Only required if you do not work in Rstudio).
• Install Rmisc to perform some basic statistical computations (e.g., calculate mean, median, sd, se, ci).
• Install gghalves from their GitHub repository since it features the newest options as opposed to the version on CRAN. To install from GitHub, the devtools package needs to be installed first. gghalves is a ggplot2 extension for easy plotting of half-half geom combinations. Think half boxplot and half jitterplot, or half violinplot and half dotplot.

Package references

• plyr - Wickham, 2019
• lattice - Sarkar, 2019
• rmarkdown - Allaire & Xie, 2020
• ggplot2 - Wickham et al., 2020
• dplyr - Wickham et al., 2020
• readr - Wickham et al., 2020
• Rmisc - Hope, 2013
• devtools - Wickham & Hester, 2020
• gghalves - Tiedemann, 2020

Install packages

packages <- c("plyr", "lattice", "ggplot2", "dplyr", "readr", "rmarkdown","Rmisc")
if (length(setdiff(packages, rownames(installed.packages()))) > 0) {
  install.packages(setdiff(packages, rownames(installed.packages())))  
}
if (!require(devtools)) {
    install.packages("devtools")
}
devtools::install_github('erocoar/gghalves')

Load packages

library("plyr")
library("lattice")
library("ggplot2")
library("dplyr")
library("readr")
library("rmarkdown")
library("Rmisc")
library("devtools")
library("gghalves")
# width and height variables for saved plots
w = 6
h = 4
# Define limits of y-axis
y_lim_min = 4
y_lim_max = 7.5
# Make the figure folder if it doesn't exist yet
dir.create('../tutorial_R/figs_repmes/', showWarnings = FALSE)

For this tutorial, we make use of the package gghalves.

The main idea behind gghalves is that standard geom's aggregate data e.g.,geom_boxplot, geom_violin and geom_dotplot who all tend to be an approximation of symmetry. Given that the space to display information is limited, we can make better use of it by cutting the geoms in half and displaying additional geoms that e.g. give information about the sample size Tiedemann, 2020.

Figure 1

For this example, we make use of the iris dataset, which is freely available in R.

We manipulate the dataset by creating two variables (1) before and (2) after. Specifiy the variable n and create a dataframe d which includes the variables y, x and id.

before = iris$Sepal.Length[1:50]
after = iris$Sepal.Length[51:100]
n <- length(before)
d <- data.frame(y = c(before, after),
                x = rep(c(1,2), each=n),
                id = factor(rep(1:n,2)))

Let's create a first very basic figure only showing the individual datapoints.

f1 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(aes(x=x), color = "magenta", size = 1.5) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 1: Repeated measures individual datapoints') +
   theme_classic() +
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f1
ggsave('../tutorial_R/figs_repmes/figure1.png', width = w, height = h)

Figure 2

Let's again create a simple figure, but now with the datapoints connected (i.e., intra-individual trends).

f2 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(aes(x=x), color = "magenta", size = 1.5) +
   geom_line(aes(x=x, group=id), color = 'lightgray') +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 2: Repeated measures with connecting lines') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f2
ggsave('../tutorial_R/figs_repmes/figure2.png', width = w, height = h)

Figure 3

Let's add some jitter to avoid that datapoints overlap.

set.seed(321)
d$xj <- jitter(d$x, amount=.09)

Now we create the the figure again, including jitter.

f3 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(aes(x=xj), color = "magenta", size = 1.5) +
   geom_line(aes(x=xj, group=id), color = 'lightgray') +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 3: Repeated measures with jitter and connections') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f3
ggsave('../tutorial_R/figs_repmes/figure3.png', width = w, height = h)

Figure 4

Produce another figure, but now defined by different colors for Before and After. Just for illustrative purposes, alpha was used in geom_line.

f4 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x=="1"), aes(x=xj), color = 'dodgerblue', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(x=="2"), aes(x=xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_line(aes(x=xj, group=id), color = 'lightgray', alpha = .3) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 4: Repeated measures with jittered datapoints and connections') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f4
ggsave('../tutorial_R/figs_repmes/figure4.png', width = w, height = h)

Figure 5

Let's add box- and violinplots to create raincloud like plots.

f5 <- ggplot(data = d, aes(y = y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x =="1"), aes(x = xj), color = 'dodgerblue', size = 1.5, 
              alpha = .6) +
   geom_point(data = d %>% filter(x =="2"), aes(x = xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_line(aes(x = xj, group = id), color = 'lightgray', alpha = .3) +

   geom_half_boxplot(
     data = d %>% filter(x=="1"), aes(x=x, y = y), position = position_nudge(x = -.25), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'dodgerblue') +

   geom_half_boxplot(
     data = d %>% filter(x=="2"), aes(x=x, y = y), position = position_nudge(x = .15), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'darkorange') +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = -.3), 
     side = "l", fill = 'dodgerblue') +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = .3), 
     side = "r", fill = "darkorange") +


   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 5: Repeated measures with box- and violin plots') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f5
ggsave('../tutorial_R/figs_repmes/figure5.png', width = w, height = h)

Figure 5b

f5b <- ggplot(data = d, aes(y = y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x =="1"), aes(x = xj), color = 'dodgerblue', size = 1.5, 
              alpha = .5) +
   geom_point(data = d %>% filter(x =="2"), aes(x = xj), color = 'darkorange', size = 1.5, 
              alpha = .5) +
   geom_line(aes(x = xj, group = id), color = 'lightgray', alpha = .5) +

   geom_half_boxplot(
     data = d %>% filter(x=="1"), aes(x=x, y = y), position = position_nudge(x = -.25), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = TRUE, width = .1, 
     fill = 'dodgerblue', alpha = .5) +

   geom_half_boxplot(
     data = d %>% filter(x=="2"), aes(x=x, y = y), position = position_nudge(x = .15), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = TRUE, width = .1, 
     fill = 'darkorange', alpha = .5) +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = 1.3), 
     side = "r", fill = 'dodgerblue', alpha = .5, trim = FALSE) +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = .3), 
     side = "r", fill = "darkorange", alpha = .5, trim = FALSE) +


   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 5b: Repeated measures with box- and violin plots') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f5b
ggsave('../tutorial_R/figs_repmes/figure5b.png', width = w, height = h)

f5c <- ggplot(data = d, aes(y = y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x =="1"), aes(x = xj), color = 'dodgerblue', size = 1.5, 
              alpha = .5) +
   geom_point(data = d %>% filter(x =="2"), aes(x = xj), color = 'darkorange', size = 1.5, 
              alpha = .5) +
   geom_line(aes(x = xj, group = id), color = 'lightgray', alpha = .5) +

   geom_half_boxplot(
     data = d %>% filter(x=="1"), aes(x=x, y = y), position = position_nudge(x = 1.21), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = TRUE, width = .1, 
     fill = 'dodgerblue', alpha = .5) +

   geom_half_boxplot(
     data = d %>% filter(x=="2"), aes(x=x, y = y), position = position_nudge(x = .15), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = TRUE, width = .1, 
     fill = 'darkorange', alpha = .5) +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = 1.3), 
     side = "r", fill = 'dodgerblue', alpha = .5, trim = FALSE) +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = .3), 
     side = "r", fill = "darkorange", alpha = .5, trim = FALSE) +


   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 5c: Repeated measures with box- and violin plots') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f5c
ggsave('../tutorial_R/figs_repmes/figure5c.png', width = w, height = h)

Note

A potential downside of the current approach is that it uses a lot of filtering and a lot of geoms.

Add descriptive statistics

• Create a dataframe including:
• Mean
• Median
• Standard deviation
• Standard error
• Confidence interval (95 %)

score_mean_1 <- mean(d$y[1:50])
score_mean_2 <- mean(d$y[51:100])
score_median1 <- median(d$y[1:50])
score_median2 <- median(d$y[51:100])
score_sd_1 <- sd(d$y[1:50])
score_sd_2 <- sd(d$y[51:100])
score_se_1 <- score_sd_1/sqrt(n) #-> adjust your n
score_se_2 <- score_sd_2/sqrt(n) #-> adjust your n
score_ci_1 <- CI(d$y[1:50], ci = 0.95)
score_ci_2 <- CI(d$y[51:100], ci = 0.95)
#Create data frame with 2 rows and 7 columns containing the descriptives
group <- c("x", "z")
N <- c(50, 50)
score_mean <- c(score_mean_1, score_mean_2)
score_median <- c(score_median1, score_median2)
sd <- c(score_sd_1, score_sd_2)
se <- c(score_se_1, score_se_2)
ci <- c((score_ci_1[1] - score_ci_1[3]), (score_ci_2[1] - score_ci_2[3]))
#Create the dataframe
summary_df <- data.frame(group, N, score_mean, score_median, sd, se, ci)

Figure 6

Let's add the items calculated in summary_df to the figure

f6 <- ggplot(data = d, aes(y = y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x =="1"), aes(x = xj), color = 'dodgerblue', size = 1.5, 
              alpha = .6) +
   geom_point(data = d %>% filter(x =="2"), aes(x = xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_line(aes(x = xj, group = id), color = 'lightgray', alpha = .3) +

   geom_half_boxplot(
     data = d %>% filter(x=="1"), aes(x=x, y = y), position = position_nudge(x = -.28), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'dodgerblue') +

   geom_half_boxplot(
     data = d %>% filter(x=="2"), aes(x=x, y = y), position = position_nudge(x = .18), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'darkorange') +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = -.3), 
     side = "l", fill = 'dodgerblue') +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = .3), 
     side = "r", fill = "darkorange") +

   geom_point(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1]), 
     position = position_nudge(x = -.13), color = "dodgerblue", alpha = .6, size = 1.5) +

   geom_errorbar(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1], 
     ymin = score_mean[1]-ci[1], ymax = score_mean[1]+ci[1]), 
     position = position_nudge(-.13), 
     color = "dodgerblue", width = 0.05, size = 0.4, alpha = .5) + 

   geom_point(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2]), 
    position = position_nudge(x = .13), color = "darkorange", alpha = .6, size = 1.5)+ 

   geom_errorbar(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2], 
    ymin = score_mean[2]-ci[2], 
    ymax = score_mean[2]+ci[2]), position = position_nudge(.13), color = "darkorange", 
    width = 0.05, size = 0.4, alpha = .5) +


   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 6: Repeated measures with box- and violin plots') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f6
ggsave('../tutorial_R/figs_repmes/figure6.png', width = w, height = h)

Figure 7

Optionally you could add a line between the two means. Define the x-coordinates where the line needs to be drawn. Note these coordinates can be calculated as 1 - position_nudge() for the first variable and 2 + position_nudge() for the second variable, which in our case is 1 - .13 = .87 and 2 + .13 = 2.13.

x_tick_means <- c(.87, 2.13)
f7 <- ggplot(data = d, aes(y = y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x =="1"), aes(x = xj), color = 'dodgerblue', size = 1.5, 
              alpha = .6) +
   geom_point(data = d %>% filter(x =="2"), aes(x = xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_line(aes(x = xj, group = id), color = 'lightgray', alpha = .3) +

   geom_half_boxplot(
     data = d %>% filter(x=="1"), aes(x=x, y = y), position = position_nudge(x = -.28), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'dodgerblue') +

   geom_half_boxplot(
     data = d %>% filter(x=="2"), aes(x=x, y = y), position = position_nudge(x = .18), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'darkorange') +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = -.3), 
     side = "l", fill = 'dodgerblue') +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = .3), 
     side = "r", fill = "darkorange") +

   geom_point(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1]), 
     position = position_nudge(x = -.13), color = "dodgerblue", alpha = .6, size = 1.5) +

   geom_errorbar(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1], 
     ymin = score_mean[1]-ci[1], ymax = score_mean[1]+ci[1]), 
     position = position_nudge(-.13), 
     color = "dodgerblue", width = 0.05, size = 0.4, alpha = .6) + 

   geom_point(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2]), 
    position = position_nudge(x = .13), color = "darkorange", alpha = .6, size = 1.5)+ 

   geom_errorbar(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2], 
    ymin = score_mean[2]-ci[2], 
    ymax = score_mean[2]+ci[2]), position = position_nudge(.13), color = "darkorange", 
    width = 0.05, size = 0.4, alpha = .6) +

  #Add a line connecting the two means
   geom_line(data = summary_df, aes(x = x_tick_means, y = score_mean), color = 'gray', 
    size = 1) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2), labels=c("Before", "After"), limits=c(0, 3)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 7: Repeated measures with box- and violin plots and means + ci ') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f7
ggsave('../tutorial_R/figs_repmes/figure7.png', width = w, height = h)

2 x 2 repeated measures rainclouds in a #butterfly fashion

As a last step, let's create these figures for a 2 x 2 repeated measures study.

Define an additional variable z which has two categories 3 and 4 and create a second jittered variable.

before = iris$Sepal.Length[1:50] 
after = iris$Sepal.Length[51:100]
n <- length(before) 
d <- data.frame(y = c(before, after),
               x = rep(c(1,2), each=n),
               z = rep(c(3,4), each=n), 
               id = factor(rep(1:n,2))) 
set.seed(321)
d$xj <- jitter(d$x, amount = .09) 
d$xj_2 <- jitter(d$z, amount = .09)

f8 <- ggplot(data = d, aes(y = y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x =="1"), aes(x = xj), color = 'dodgerblue', size = 1.5, 
              alpha = .6) +
   geom_point(data = d %>% filter(x =="2"), aes(x = xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_point(data = d %>% filter(z =="3"), aes(x = xj_2), color = 'dodgerblue', size = 1.5, 
              alpha = .6) + 
   geom_point(data = d %>% filter(z =="4"), aes(x = xj_2), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_line(aes(x = xj, group = id), color = 'lightgray', alpha = .3) +
   geom_line(aes(x = xj_2, group = id), color = 'lightgray', alpha = .3) +

   geom_half_boxplot(
     data = d %>% filter(x=="1"), aes(x=x, y = y), position = position_nudge(x = -.35), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'dodgerblue', alpha = .6) +

   geom_half_boxplot(
     data = d %>% filter(x=="2"), aes(x=x, y = y), position = position_nudge(x = -1.16), 
     side = "l",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'darkorange', alpha = .6) +

   geom_half_boxplot(
     data = d %>% filter(z=="3"), aes(x=z, y = y), position = position_nudge(x = 1.3), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'dodgerblue', alpha = .6) +

   geom_half_boxplot(
     data = d %>% filter(z=="4"), aes(x=z, y = y), position = position_nudge(x = .2), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'darkorange', alpha = .6) +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = -.40), 
     side = "l", fill = 'dodgerblue', alpha = .6) +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = -1.40), 
     side = "l", fill = "darkorange", alpha = .6) +

   geom_half_violin(
     data = d %>% filter(z=="3"),aes(x = z, y = y), position = position_nudge(x = 1.45), 
     side = "r", fill = 'dodgerblue', alpha = .6) +

   geom_half_violin(
     data = d %>% filter(z=="4"),aes(x = z, y = y), position = position_nudge(x = .45), 
     side = "r", fill = "darkorange", alpha = .6) +

   geom_point(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1]), 
     position = position_nudge(x = -.13), color = "dodgerblue", alpha = .6, size = 1.5) +

   geom_errorbar(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1], 
     ymin = score_mean[1]-ci[1], ymax = score_mean[1]+ci[1]), 
     position = position_nudge(-.13), 
     color = "dodgerblue", width = 0.05, size = 0.4, alpha = .6) + 

   geom_point(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2]), 
    position = position_nudge(x = -1.1), color = "darkorange", alpha = .6, size = 1.5)+ 

   geom_errorbar(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2], 
    ymin = score_mean[2]-ci[2], 
    ymax = score_mean[2]+ci[2]), position = position_nudge(x = -1.1), color = "darkorange", 
    width = 0.05, size = 0.4, alpha = .6) +
   geom_point(data = d %>% filter(z=="3"), aes(x = z, y = score_mean[1]), 
    position = position_nudge(x = 1.15), color = "dodgerblue", alpha = .5) +

   geom_errorbar(data = d %>% filter(z=="3"), aes(x = z, y = score_mean[1], 
    ymin = score_mean[1]-ci[1], 
    ymax = score_mean[1]+ci[1]), position = position_nudge(1.15), 
    color = "dodgerblue", width = 0.05, size = 0.4, alpha = .5)+ 
   geom_point(data = d %>% filter(z=="4"), aes(x = z, y = score_mean[2]), 
    position = position_nudge(x = .15), color = "darkorange", alpha = .5)+
   geom_errorbar(data = d %>% filter(z=="4"), aes(x = z, y = score_mean[2], 
    ymin = score_mean[2]-ci[2], ymax = score_mean[2]+ci[2]), position = position_nudge(.15), 
    color = "darkorange", width = 0.05, size = 0.4, alpha = .5)+


   #Define additional settings
   scale_x_continuous(breaks=c(1,2,3,4), labels=c("Before", "After","Before", "After"), 
                      limits=c(0, 5))+
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 8: 2 x 2 Repeated measures with box- and violin plots') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f8
ggsave('../tutorial_R/figs_repmes/figure8.png', width = w, height = h)

Figure 9

Finally, let's add lines that connect the means of each group.

#First we must again define the x-coordinates of the means.
x_tick_means_x <- c(.87, 2.13) #same as above
x_tick_means_z <- c(2.87, 4.13) #just add 2 for each tick
f9 <- ggplot(data = d, aes(y = y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x =="1"), aes(x = xj), color = 'dodgerblue', size = 1.5, 
              alpha = .6) +
   geom_point(data = d %>% filter(x =="2"), aes(x = xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_point(data = d %>% filter(z =="3"), aes(x = xj_2), color = 'dodgerblue', size = 1.5, 
              alpha = .6) + 
   geom_point(data = d %>% filter(z =="4"), aes(x = xj_2), color = 'darkorange', size = 1.5, 
              alpha = .6) +
   geom_line(aes(x = xj, group = id), color = 'lightgray', alpha = .3) +
   geom_line(aes(x = xj_2, group = id), color = 'lightgray', alpha = .3) +

   geom_half_boxplot(
     data = d %>% filter(x=="1"), aes(x=x, y = y), position = position_nudge(x = -.35), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'dodgerblue', alpha = .6) +

   geom_half_boxplot(
     data = d %>% filter(x=="2"), aes(x=x, y = y), position = position_nudge(x = -1.16), 
     side = "l",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'darkorange', alpha = .6) +

   geom_half_boxplot(
     data = d %>% filter(z=="3"), aes(x=z, y = y), position = position_nudge(x = 1.3), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'dodgerblue', alpha = .6) +

   geom_half_boxplot(
     data = d %>% filter(z=="4"), aes(x=z, y = y), position = position_nudge(x = .2), 
     side = "r",outlier.shape = NA, center = TRUE, errorbar.draw = FALSE, width = .2, 
     fill = 'darkorange', alpha = .6) +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = -.40), 
     side = "l", fill = 'dodgerblue', alpha = .6) +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = -1.40), 
     side = "l", fill = "darkorange", alpha = .6) +

   geom_half_violin(
     data = d %>% filter(z=="3"),aes(x = z, y = y), position = position_nudge(x = 1.45), 
     side = "r", fill = 'dodgerblue', alpha = .6) +

   geom_half_violin(
     data = d %>% filter(z=="4"),aes(x = z, y = y), position = position_nudge(x = .45), 
     side = "r", fill = "darkorange", alpha = .6) +

   geom_point(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1]), 
     position = position_nudge(x = -.13), color = "dodgerblue", alpha = .6, size = 1.5) +

   geom_errorbar(data = d %>% filter(x=="1"), aes(x = x, y = score_mean[1], 
     ymin = score_mean[1]-ci[1], ymax = score_mean[1]+ci[1]), 
     position = position_nudge(-.13), 
     color = "dodgerblue", width = 0.05, size = 0.4, alpha = .6) + 

   geom_point(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2]), 
    position = position_nudge(x = .13), color = "darkorange", alpha = .6, size = 1.5)+ 

   geom_errorbar(data = d %>% filter(x=="2"), aes(x = x, y = score_mean[2], 
    ymin = score_mean[2]-ci[2], 
    ymax = score_mean[2]+ci[2]), position = position_nudge(x = .13), color = "darkorange", 
    width = 0.05, size = 0.4, alpha = .6) +
   geom_point(data = d %>% filter(z=="3"), aes(x = z, y = score_mean[1]), 
    position = position_nudge(x = -.13), color = "dodgerblue", alpha = .5) +

   geom_errorbar(data = d %>% filter(z=="3"), aes(x = z, y = score_mean[1], 
    ymin = score_mean[1]-ci[1], 
    ymax = score_mean[1]+ci[1]), position = position_nudge(-.13), 
    color = "dodgerblue", width = 0.05, size = 0.4, alpha = .5)+ 
   geom_point(data = d %>% filter(z=="4"), aes(x = z, y = score_mean[2]), 
    position = position_nudge(x = .13), color = "darkorange", alpha = .5)+
   geom_errorbar(data = d %>% filter(z=="4"), aes(x = z, y = score_mean[2], 
    ymin = score_mean[2]-ci[2], ymax = score_mean[2]+ci[2]), 
    position = position_nudge(.13), 
    color = "darkorange", width = 0.05, size = 0.4, alpha = .5)+

   #Add lines connecting the two means
   geom_line(data = summary_df, aes(x = x_tick_means_x, y = score_mean), 
             color = 'gray', size = 1) +
   geom_line(data = summary_df, aes(x = x_tick_means_z, y = score_mean), 
             color = 'gray', size = 1) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2,3,4), labels=c("Before", "After","Before", "After"), 
                      limits=c(0, 5))+
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 9: 2 x 2 Repeated measures with box- and violin plots') +
   theme_classic()+
   coord_cartesian(ylim=c(y_lim_min, y_lim_max))
f9
ggsave('../tutorial_R/figs_repmes/figure9.png', width = w, height = h)

Figure 10

Here, we add an additional variable 'during' if you want to show the effect measured over 3 timepoints.

before = iris$Sepal.Length[1:50]
during = iris$Sepal.Length[51:100]
after = iris$Sepal.Length[1:50]
n <- length(before) 
d <- data.frame(y = c(before, during, after),
               x = rep(c(1,2,3), each=n),
               id = factor(rep(1:n,3)))
set.seed(321)
d$xj <- jitter(d$x, amount = .09) 
f10 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x=="1"), aes(x=xj), color = 'dodgerblue', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(x=="2"), aes(x=xj), color = 'darkgreen', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(x=="3"), aes(x=xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +

   geom_line(aes(x=xj, group=id), color = 'lightgray', alpha = .3) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2,3), labels=c("Before", "During", "After"), limits=c(0.5, 3.5)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 10: Repeated measures with jittered datapoints and connections') +
   theme_classic()+
   coord_cartesian(ylim=c(3.5, 7.5))
f10
ggsave('../tutorial_R/figs_repmes/figure10.png', width = w, height = h)

before = iris$Sepal.Width[1:50]
during = iris$Sepal.Length[51:100]
after = iris$Sepal.Length[1:50]
n <- length(before) 
d <- data.frame(y = c(before, during, after),
               x = rep(c(1,2,3), each=n),
               id = factor(rep(1:n,3)))
set.seed(321)
d$xj <- jitter(d$x, amount = .09) 
f11 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x=="1"), aes(x=xj), color = 'dodgerblue', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(x=="2"), aes(x=xj), color = 'darkgreen', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(x=="3"), aes(x=xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +

   geom_line(aes(x=xj, group=id), color = 'lightgray', alpha = .3) +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = 2.2), 
     side = "r", fill = 'dodgerblue', alpha = .5, color = "dodgerblue", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = 1.2), 
     side = "r", fill = "darkgreen", alpha = .5, color = "darkgreen", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(x=="3"),aes(x = x, y = y), position = position_nudge(x = 0.2), 
     side = "r", fill = "darkorange", alpha = .5, color = "darkorange", trim = TRUE) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2,3), labels=c("Before", "During", "After"), limits=c(0.5, 4)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 11: Repeated measures with jittered datapoints and connections') +
   theme_classic()+
   coord_cartesian(ylim=c(2.3, 7.2))
f11
ggsave('../tutorial_R/figs_repmes/figure11.png', width = w, height = h)

f12 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x=="1"), aes(x=xj), color = 'dodgerblue', size = 1.5,
              alpha = .8) +
   geom_point(data = d %>% filter(x=="2"), aes(x=xj), color = 'darkgreen', size = 1.5,
              alpha = .8) +
   geom_point(data = d %>% filter(x=="3"), aes(x=xj), color = 'darkorange', size = 1.5, 
              alpha = .8) +

   geom_line(aes(x=xj, group=id), color = 'lightgray', alpha = .3) +

   geom_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = 0),
     fill = 'dodgerblue', alpha = .1, color = "white", trim = TRUE) +

   geom_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = 0),
     fill = "darkgreen", alpha = .1, color = "white", trim = FALSE) +

   geom_violin(
     data = d %>% filter(x=="3"),aes(x = x, y = y), position = position_nudge(x = 0),
     fill = "darkorange", alpha = .1, color = "white", trim = TRUE) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2,3), labels=c("Before", "During", "After"), limits=c(0.5, 4)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 12: Repeated measures with jittered datapoints and connections') +
   theme_classic()+
   coord_cartesian(ylim=c(2.3, 7.2))
f12
ggsave('../tutorial_R/figs_repmes/figure12.png', width = w, height = h)

before = iris$Sepal.Width[1:50]
during = iris$Sepal.Length[51:100]
after = iris$Sepal.Length[1:50]
n <- length(before) 
d <- data.frame(y = c(before, during, after),
               x = rep(c(1,2,3), each=n),
               z = rep(c(4,5,6), each=n), 
               id = factor(rep(1:n,3)))
set.seed(321)
d$xj <- jitter(d$x, amount = .09) 
d$xj_2 <- jitter(d$z, amount = .09)
#d$xj_3 <- jitter(d$a, amount = .09)
f13 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x=="1"), aes(x=xj), color = 'dodgerblue', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(x=="2"), aes(x=xj), color = 'darkgreen', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(x=="3"), aes(x=xj), color = 'darkorange', size = 1.5, 
              alpha = .6) +

    #Add geom_() objects
   geom_point(data = d %>% filter(z=="4"), aes(x=xj_2), color = 'dodgerblue', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(z=="5"), aes(x=xj_2), color = 'darkgreen', size = 1.5,
              alpha = .6) +
   geom_point(data = d %>% filter(z=="6"), aes(x=xj_2), color = 'darkorange', size = 1.5, 
              alpha = .6) +

   geom_line(aes(x=xj, group=id), color = 'lightgray', alpha = .3) +
   geom_line(aes(x=xj_2, group=id), color = 'lightgray', alpha = .3) +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = -0.2), 
     side = "l", fill = 'dodgerblue', alpha = .5, color = "dodgerblue", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = -1.2), 
     side = "l", fill = "darkgreen", alpha = .5, color = "darkgreen", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(x=="3"),aes(x = x, y = y), position = position_nudge(x = -2.2), 
     side = "l", fill = "darkorange", alpha = .5, color = "darkorange", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(z=="4"),aes(x = x, y = y), position = position_nudge(x = 5.2), 
     side = "r", fill = 'dodgerblue', alpha = .5, color = "dodgerblue", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(z=="5"),aes(x = x, y = y), position = position_nudge(x = 4.2), 
     side = "r", fill = "darkgreen", alpha = .5, color = "darkgreen", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(z=="6"),aes(x = x, y = y), position = position_nudge(x = 3.2), 
     side = "r", fill = "darkorange", alpha = .5, color = "darkorange", trim = TRUE) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2,3,4,5,6), labels=c("Before", "During", "After", "Before", "During", "After"), limits=c(0, 7)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 13: Repeated measures with jittered datapoints and connections') +
   theme_classic()+
   coord_cartesian(ylim=c(2.3, 7.2))
f13
ggsave('../tutorial_R/figs_repmes/figure13.png', width = w, height = h)

f14 <- ggplot(data=d, aes(y=y)) +

   #Add geom_() objects
   geom_point(data = d %>% filter(x=="1"), aes(x=xj), color = 'dodgerblue', size = 1.5,
              alpha = .8) +
   geom_point(data = d %>% filter(x=="2"), aes(x=xj), color = 'darkgreen', size = 1.5,
              alpha = .8) +
   geom_point(data = d %>% filter(x=="3"), aes(x=xj), color = 'darkorange', size = 1.5, 
              alpha = .8) +

    #Add geom_() objects
   geom_point(data = d %>% filter(z=="4"), aes(x=xj_2), color = 'dodgerblue', size = 1.5,
              alpha = .8) +
   geom_point(data = d %>% filter(z=="5"), aes(x=xj_2), color = 'darkgreen', size = 1.5,
              alpha = .8) +
   geom_point(data = d %>% filter(z=="6"), aes(x=xj_2), color = 'darkorange', size = 1.5, 
              alpha = .8) +

   geom_line(aes(x=xj, group=id), color = 'lightgray', alpha = .3) +
   geom_line(aes(x=xj_2, group=id), color = 'lightgray', alpha = .3) +

   geom_half_violin(
     data = d %>% filter(x=="1"),aes(x = x, y = y), position = position_nudge(x = 0), 
     side = "r", fill = 'dodgerblue', alpha = .15, color = "lightgray", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(x=="2"),aes(x = x, y = y), position = position_nudge(x = 0), 
     side = "r", fill = "darkgreen", alpha = .15, color = "lightgray", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(x=="3"),aes(x = x, y = y), position = position_nudge(x = 0), 
     side = "r", fill = "darkorange", alpha = .15, color = "lightgray", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(z=="4"),aes(x = x, y = y), position = position_nudge(x = 3), 
     side = "r", fill = 'dodgerblue', alpha = .15, color = "lightgray", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(z=="5"),aes(x = x, y = y), position = position_nudge(x = 3), 
     side = "r", fill = "darkgreen", alpha = .15, color = "lightgray", trim = TRUE) +

   geom_half_violin(
     data = d %>% filter(z=="6"),aes(x = x, y = y), position = position_nudge(x = 3), 
     side = "r", fill = "darkorange", alpha = .15, color = "lightgray", trim = TRUE) +

   #Define additional settings
   scale_x_continuous(breaks=c(1,2,3,4,5,6), labels=c("Before", "During", "After", "Before", "During", "After"), limits=c(0, 7)) +
   xlab("Condition") + ylab("Value") +
   ggtitle('Figure 14: Repeated measures with jittered datapoints and connections') +
   theme_classic()+
   coord_cartesian(ylim=c(2.3, 7.2))
f14
ggsave('../tutorial_R/figs_repmes/figure14.png', width = w, height = h)

General remarks / tips

• To be more flexible in assigning labels to your figures, the ggtext package by Wilke, 2020 might be worthwile .
• If you would like to be flexible in plotting multiple figures next to each other, check-out the patchwork package by Pedersen, 2020.
• If you want to save your figures in a high-quality manner (> GB) for e.g., publications, you could save your figure with a .tiff extension and add dpi= as used in the following line of code:

ggsave("figure.tiff", height=h, width=w, units='in', dpi=600)

• If for some reason your code does not work due to e.g., an update in a package, you could use the following line of code to unload and load that package again.

if("package_name" %in% (.packages())){
  detach('package:package_name', unload=TRUE)}
library("package_name")

You have reached the end of this document.

I hope you'll be able to use this tutorial to create more open-visualizations for your research!

If you use this tutorial, please cite it in your work.

open-visualizations for repeated measures in R by Jordy van Langen

jorvlan/openvis documentation built on Dec. 6, 2020, 3:06 a.m.