R in Python (Using R ggplot to make visualizations in Jupyter Notebook)

I have been using Python as my daily tool and got a chance to work with R during my Spring 2017 semester at Columbia University. I appreciate R for its statistical tools and amazing visualization libraries like ggplot. For the final project, I wanted to use Python again but also wanted to use R visualization tools.

My online search led me to a python package called rpy2 and I will show how we can use it to integrate R in Jupyter Notebooks via Cell Magics.

My preferred way to set up Python Environments

In the past I liked using virtualenv and virtualenvwrapper, but have recently shifted to using Conda and use it everyday in my workflow. A general walkthrough follows :

Create the virtual environment 

conda create -n py_R python=3.6

Install the pre-requistie packages 

conda install rpy2
conda install jupyter
conda install pandas

Imports for the Code

In [1]:
## Lets try out some code
import pandas as pd         # do all processing in pandas, and convert Pandas DataFrame to R DataFrame

## imports required from rpy2
from rpy2.robjects import pandas2ri   
pandas2ri.activate()
%reload_ext rpy2.ipython
In [2]:
## HTML changes to align the images to the center of the screen

from IPython.core.display import HTML
HTML("""

""")
Out[2]:

Pandas DataFrame : Just initialize a data frame

In [9]:
df = pd.DataFrame({'Letter': ['a', 'a', 'a', 'b','b', 'b', 'c', 'c','c'],
                   'X': [4, 3, 5, 2, 1, 7, 7, 5, 9],
                   'Y': [0, 4, 3, 6, 7, 10, 11, 9, 13],
                   'Z': [1, 2, 3, 1, 2, 3, 1, 2, 3]})
In [10]:
df
Out[10]:
Letter X Y Z
0 a 4 0 1
1 a 3 4 2
2 a 5 3 3
3 b 2 6 1
4 b 1 7 2
5 b 7 10 3
6 c 7 11 1
7 c 5 9 2
8 c 9 13 3

Using rpy2 is as simple as using cell magic R

You can also specify the input pandas dataframe to R cell using %%R with -i with the height and width specified after -w and -h arguments and the units specified using the -u argument

In [12]:
%%R -i df -w 900 -h 480 -u px

## Everything in here is ** R ** (magic)

print(df) # Tada, df is now an R dataframe

library("ggplot2") # If this line does not work, make sure you have R installed on your laptop

ggplot(data = df) + geom_point(aes(x = X, y = Y, color = Letter, size = Z))

  Letter X  Y Z
0      a 4  0 1
1      a 3  4 2
2      a 5  3 3
3      b 2  6 1
4      b 1  7 2
5      b 7 10 3
6      c 7 11 1
7      c 5  9 2
8      c 9 13 3

Comments

comments powered by Disqus