Integrating R with LaTeX using knitr

A comprehensive guide to integrating R code and LaTeX documents using the knitr package, enabling dynamic report generation and reproducible research.
R
LaTeX
Statistics
Data Science
Author

Anant Kumar

Published

February 19, 2026

Introduction

I had heard about R but never really got into it until I started with a course on Statistical Computing. I needed to create reports that required me to show the output of R codes. I have been working in the type setting system LaTex and my instinct was to put the screenshot as an image in the source file. Soon I was caught in a cycle: You run a simulation in R, copy the graph into your LaTeX document, and then realize you made a small error in the initial parameters. You fix the code, re-run it, copy the graph again and so on. That’s when I discovered knitr, a powerful package in R that allows you to integrate R code with LaTeX documents seamlessly.

In this post, I will share my experience and provide a comprehensive guide on how to use knitr to create dynamic reports that combine R code and LaTeX. This approach not only saves time but also ensures that your reports are reproducible and up-to-date with the latest data and analyses.

What is knitr?

knitr is an R package that allows you to embed R code chunks directly into a document. While many of us use it with Markdown (like in this blog!), it was originally designed to supercharge LaTex.

Instead of writing a .tex file, we write an .Rnw (R NoWeb) file. When we compile it, knitr executes the R code, grabs the results (plots, tables, numbers), and produces a standard .tex file that is then compiled into a PDF.

Installation & Setup

Before we start, ensure that you have a working LaTex distribution (like TexLive or MikTex) and R installed. My working machine is a desktop with Ubuntu 24.04 LTS version. I have TexLive distribution and all the demos pertain to this setup. Yours might be different so set up accordingly.

Open your R console (in case, you are installing it system wide, you might have to start your R console using sudo R)or RStudio and install the knitr package:

install.packages("knitr")

If you want to create tables that play nicely with LaTex, you might also want to install xtable:

install.packages("xtable")

Creating .Rnw file

An .Rnw file looks like a regular .tex file but with a special power: Code Chunks. These code chunks are the actual R codes. Here is a minimum working example of what an .Rnw file looks like:

\documentclass{article}
\usepackage{graphicx}
\usepackage{hyperref}

\title{Analysis of Randomness}
\author{Anant Kumar}

\begin{document}

\maketitle

\section{Introduction}
This is a document generated using \textbf{knitr}. 
Below, we generate some random numbers.

% This is a code chunk
<<chunk-label, echo=TRUE, fig.width=6, fig.height=4>>=
# R code goes here
set.seed(123)
x <- rnorm(100)
mean_x <- mean(x)
@

The mean of our generated data is \Sexpr{mean_x}.

\end{document}

Syntax Breakdown

1. The Code Chunk

Chunks are where the magic happens. They start with <<options>>= and end with @.

<<my-plot, echo=FALSE>>=
plot(x)
@

2. Inline R code (\Sexpr)

This is actually a very useful feature. Instead of typing “The value is 5.32” and having to update it later, you write:

The value is \Sexpr{mean(x)}.

When compiled, knitr calculates the mean and inserts the number directly into the sentence.

Customizing Chunks

knitr provides granular control over the output that it produces. Here are the most common options you will use in <<>>= header:

  • echo=FALSE: Hides the code but shows the results (if you just wanted to report the output).

  • eval=FALSE: Shows the code but doesn’t run it (useful for tutorials).

  • include=FALSE: Runs the code but hides everything (good for loading packages or setting up data).

  • results='asis': Critical for tables. It tells LaTeX to treat the R output as raw LaTeX text.

  • fig.cap="My Caption": Automatically adds a figure caption.

Generating Tables

In standard LaTeX, creating tables is tedious. With R and the xtable package, it becomes dynamic.

<<table-example, results='asis', echo=FALSE>>=
library(xtable)
data(iris)
# Create a LaTeX table for the first 5 rows of the iris dataset
print(xtable(head(iris), caption="Iris Data Preview"))
@

Because we used results='asis', R outputs the raw LaTeX code for the table (\begin{table} ... \end{table}), which is then rendered perfectly in the PDF.

Compiling the Document

Once done editing, go to the R console (ensuring that you start it in the same directory in which the document is present) and run the following R command:

library(knitr)
knit2pdf("your_document.Rnw")

This function:

  1. Knits: Converts mydocument.Rnw \(\rightarrow\) mydocument.tex.

  2. Compiles: Runs pdflatex on mydocument.tex \(\rightarrow\) mydocument.pdf.

Conclusion

Integrating R into LaTeX has saved me countless hours when preparing documents that contain R codes. It ensures that if I change a dataset or a parameter, every single graph, table, and inline number in my document updates automatically.

Give it a try for your next project report or assignment!