For each topic that we cover every week, there are numerous additional packages and techniques that we won’t have time to cover directly. This is a short list of some of those packages and techniques. I will add to this list as we go, if you have something to add to the list, let me know, or fork this repo and make a pull-request.
ggarrange function which allows you to knit multiple plots togetherHow do you control the size (width/height) of the figure, and where can you can get a file, like a .pdf or a .png of the figure?
I use knitr options to solve these problems.
When you create a new R Markdown document you will see the following code chunk underneath the yaml, at the beginning of the document. It usually looks like this:
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```This chunk is named setup, it is not printed in your output, but it controls the global setup options for the whole document. Any option you set here applies to all of the remaining code chunks you create. It’s a way of setting defaults.
Here are some helpful defaults you can add. You turn the option on with TRUE, and turn it off with FALSE.
echo=TRUE sets the default to print all remaining code blocks to the output, FALSE sets the default to not print the code blockswarning = FALSE turns off printing of warningsmessage = FALSE turns off printing of messages, these commonly occur when you load a package, where you receive a message that the package was loadedeval = FALSE sets the default to NOT evaluate the code chunk as R Code. This will not run the code block, but the code block will still print if echo=TRUEerror=TRUE normally the knit fails when there is an error in the code. If you set error=TRUE the knit will complete, and return an error message for code blocks with errors.```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE,
warning = FALSE,
message = FALSE,
eval = FALSE,
error = TRUE)
```The following setup options are useful for figure output.
fig.width = 3 sets the default width in inches for all figuresfig.height = 3 sets the default height in inches for all figuresfig.path = "myfigs/" defines folder where figure files will be saved. This will be relative to your current working directoydev = c("pdf", "png") tells knitr to output both .png, and .pdf versions of the figure. The .pdf contains vector graphics, meaning the figure can be resized without pixelization.```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE,
fig.width = 3,
fig.height = 3,
fig.path = "myfigs/",
dev = c("pdf", "png"))
```You can set the options for remaining code blocks individually. These will overrule the default options that you specify in the setup chunk.
```{r figurename, fig.width =5, fig.height =7}
# some code that generates a figure
```There are several useful (incredible even) packages for making tables in R. Here are a few:
knitr package and it’s kable function (Create tables in Latex, HTML, Markdown)xtable package, lots of functions for tables. Here are some xtable examples xtablekableExtra lots of additional table functionalitytangram a grammar of tables approachpapaja for apa-style tablesTwo things to note, tables can be difficult, and there are different output formats.
Tables in R can be difficult. For example, if you are comfortable with making tables in Excel, then R will be much more difficult by comparison. In Excel, it is easy to enter information in any cell, merge any cells, add any kind of formatting you want anywhere, just by clicking around and making the changes. In R, every detail of the table is specified by script. Some of the table packages make simple tables easy (hurray!), and some of them make complicated tables possible (also good), but not necessarilly easy.
R can output tables in many formats including HTML for the web, Latex for .pdf, and other formats (e.g., word, markdown). Sometimes (depending on the functions you are using) you will run into issues with outputting tables in different formats. Or, you will have to take steps to ensure that you are outputting the table in the format that you want.
kable() is a function inside the knitr package. To use it you need to load the knitr library, or you can use knitr::kable(), which tells R to find kable inside the knitr package if you haven’t loaded knitr using library(knitr).
kable() is great for quickly rendering data frames as nice tables without much hassle.
df <- data.frame(A=1,
B=2,
C=3,
D=4)
knitr::kable(df)| A | B | C | D |
|---|---|---|---|
| 1 | 2 | 3 | 4 |
Look at the xtable examples document for more info. https://cran.r-project.org/web/packages/xtable/vignettes/xtableGallery.pdf
library(xtable) # load xtable
data(tli) # loads a sample data frame
# conduct an ANOVA
fm1 <- aov(tlimth ~ sex + ethnicty + grade + disadvg, data = tli)
# print the table for HTML using xtable and kable together
knitr::kable(xtable(fm1))| Df | Sum Sq | Mean Sq | F value | Pr(>F) | |
|---|---|---|---|---|---|
| sex | 1 | 75.37255 | 75.37255 | 0.3751912 | 0.5416830 |
| ethnicty | 3 | 2572.14918 | 857.38306 | 4.2679008 | 0.0071831 |
| grade | 1 | 36.30740 | 36.30740 | 0.1807318 | 0.6717271 |
| disadvg | 1 | 59.30338 | 59.30338 | 0.2952017 | 0.5882062 |
| Residuals | 93 | 18682.86749 | 200.89105 | NA | NA |
# Note this will print a table in latex for .pdf
# xtable(fm1)There are many great things about kableExtra. One great thing is that kableExtra has unique table options for html and for formatting to .pdf through latex. For example, interactive tables are possible in html, but not in .pdf. Another great thing is the ability to add rows of headers on top of each other. For example, data.frames in R only have one row of headers for the columns, kableExtra can add more on top.
library(kableExtra)
df <- data.frame(A=1,
B=2,
C=3,
D=4)
kable(df) %>%
kable_styling("striped") %>%
add_header_above(c("Group 1" = 2, "Group 2" = 2))| A | B | C | D |
|---|---|---|---|
| 1 | 2 | 3 | 4 |
This package implements a grammar of tables. Similar to the concept behind ggplot2, which implements a grammar of graphics for figures.
papaja is a package for rendering APA-style manuscripts in pdf using R Markdown. We will learn more about papaja in class. One feature of papaja is that it supports APA-style tables.