R from Python – an rpy2 tutorial

Spread the love

Rpy2 Tutorial: How to Run R in Python

In this post, we are going to answer the question can you run R in Python? Of course, the answer is yes!; using the Python package rpy2. This package offers a Python interface to R.

In this tutorial, we will learn how to use rpy2 to install r packages, and run r functions to carry out data analysis and data visualization. More specifically, we will learn how to sue the r packages r-packages ‘afex‘ and ‘emmeans‘,  usin Python, and rpy2.  Finally, we will also learn how to display R plots in Jupyter notebooks using rpy2, using two different methods.

Obviously; rpy2 requires that we have both R (version +3.2.x) and Python (versions 2.7 and 3.X) installed. There are pre-compiled binaries available for Linux and Windows (unsupported and unofficial, however).

What is rpy2?

Rpy2 is a straightforward, easy-to-use package that permit us to run R from Python. That is, this handy Python package enables us to enjoy the elegance of the Python programming  language at the same time as we get access to the rich graphical and statistical capabilities of R statistical programming environment.

How to Install rpy2

First, we start the tutorial with installing rpy2. There are two very easy ways to install rpy2.

1. Install rpy2 using pip

sudo pip install rpy2

2. Install rpy2 using conda

If you are a Windows and/or have Anaconda Python distribution installed here’s how you can install rpy2:

How to Install Rpy2

rpy2 Example: How to Call R from Python

Now when we have a working installation of rpy2, we continue the R in Python tutorial with importing the methods that we are going to use. In the following rpy2 example we are going to use ‘afex’ to do the within-subject ANOVA and ’emmeans’ to do the follow-up analysis.

import rpy2.robjects as robjects
import rpy2.robjects.packages as rpackages
from rpy2.robjects.vectors import StrVector

Before we continue with the rpy2 exampe, we also need to check whether the needed r packages are installed. In the example below we are calling r from Python to use the r package utils to install the needed r packages. The code is now updated thanks to comments on my YouTube Channel (the variable have_packages is removed. Thanks Sergey).

How to Install r packages Using rpy2

packageNames = ('afex', 'emmeans')
utils = rpackages.importr('utils')
utils.chooseCRANmirror(ind=1)

packnames_to_install = [x for x in packageNames if not rpackages.isinstalled(x)]

if len(packnames_to_install) > 0:
    utils.install_packages(StrVector(packnames_to_install))

For this tutorial we use a data set from the package Psych. In this case, we use the r-function read.table to get the data. Note how we use the class robjects from rpy2 and with a string argument we call read.table:

data = robjects.r('read.table(file =
       "http://personality-project.org/r/datasets/R.appendix3.data", header = T)')
data.head()

Repeated Measures ANOVA using rpy2

In this part of the rpy2 tutorial, we will carry out the actual analysis. In the example below we are actually using R in Python! More specifically, we are importing the r package needed to carry out our ANOVA for within-subjects design. When this is done, we will use the function aov_ez to conduct the analysis.

afex = rpackages.importr('afex')
model = afex.aov_ez('Subject', 'Recall', data, within='Valence')
print(model)

The last line above prints the results. A main effect of Valence was found.

   Effect         df  MSE          F ges p.value
1 Valence 1.15, 4.60 9.34 189.11 *** .93  < .0001
Afex ANOVA Table

Follow-up analysis

Typically we are interested in following up the main effect, and with rpy2, we can do that using the r-package ’emmeans’. First, we need to import the package and then we do a pairwise contrast and adjust for familywise error using Holm-Bonferroni correction.

emmeans = rpackages.importr('emmeans',
               robject_translations = {"recover.data.call": "recover_data_call1"})
pairwise = emmeans.emmeans(model, "Valence", contr="pairwise", adjust="holm")

That was easy, right. Now we have learned how to use R in Python! Rpy2 is relatively easy to use I don’t think it will replace learning R. That is, you will have to know some R to make use of it to call R from Python. However, if you are a Python programmer and want to use available R-scripts, it might be useful and hopefully this rpy2 tutorial have made it somewhat easier for you! Noteworthy, I am not aware of any Python implementations of rmANOVA (except for the linear-mixed effects approach maybe). In fact, that is why I learned how to use rpy2 first; to use Python, and R, to conduct the analysis. The above code examples can be found in this Jupyter Notebook.

Rpy2 Video Tutorial: Displaying R plots inline in Jupyter Notebooks

In this video we will learn how to display R plots in Jupyter Notebooks. In these two rpy2 examples we are creating a barplot (using R graphics) and a scatterplot (using ggplot2)!

R Plots in Jupyter Notebooks Using the rmagic Extension

In the last example, on how to work with R in Python we will learn an alternative method on how to display R plots in Python. Here we use the Jupyter extension rmagic.

%load_ext rpy2.ipython
%Rdevice png

In the code example above, we load the extension rmagic to enable us to run R in Python and jupyter notebooks. Next, we are setting the device to png.

After we have loaded the rmagic extension we can reproduce the first plot in the above YouTube video with fewer lines of code:

%%R
barplot(c(1,3,2,5,4), ylab="value")
Barplot

Note, we added the %%R and Jupyter will after that use rpy2 to run R functions. In the final example we are (nearly) reproducing the second plot in the above YouTube video.

%%R
require(ggplot2)
gg < - ggplot(mtcars, aes(x=wt, y=mpg)) + geom_point(aes(colour='qsec')) + theme_bw()

gg
Scatterplot

As can be seen in the code chunk above, we can also load the r package(s) we want to work with and use them much like we do in R. a

Update: In this rpy2 tutorial you learned how to do a repeated measures ANOVA with Python and R. I have now found a Python package that allows Python ANOVA for within-subjects design (i.e., Python native); see my tutorial Repeated Measures ANOVA using Python.


Spread the love

4 Comments

  1. Hi, thanks for posting this. I think you may need ‘rpackages.importr’ in place of ‘importr’ when importing the afex and lsmeans packages. That’s what worked for me anyway.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: