Data Visualization: A Python/Polars Companion

Preface

This is a Python Polars companion to Kieran Healy’s Data Visualization: A Practical Introduction. It reproduces almost all of the book’s examples using Polars for data manipulation and plotnine for plotting. While I target the second edition of Healy (2026), this companion should also work as a companion to the first edition, as many of the changes between editions relate to changes in the R code that do not affect the translation here.

A Brief Overview of Data Science using R and Python

Python and R have emerged as the dominant programming languages for data science, serving as the primary environments for statistical computing, machine learning, and data analysis (e.g., Wickham, Çetinkaya-Rundel, et al. 2023; McKinney 2022).

Data frames

Data frames are table-like (or spreadsheet-like) data structures in which each column represents a variable and each row represents an observation. The columns of a data frame may have different data types from each other.

With the benefit of some hindsight, it is clear that a core requirement for data science work is a data frame library.¹ A typical data frame library provides note the data frame data structure itself, but also functionality for working with data frames. It can be helpful to think in terms of the verbs that represent the kinds of data manipulation we want to perform on our data frames. Some of the most common things we want to do with data frames are the following:²

• filter rows based on their values
• arrange or reorder rows
• select columns
• mutate or create new variables from existing ones
• summarize many values into a smaller set of summary statistics

In addition, we likely want to combine multiple data frames and rearrange data frames from wide to long forms or vice versa.

While R comes with a built-in data frame library that handles all these tasks, a number of alternative data frame libraries have emerged over the years. The data.table library (Barrett et al. 2026) “provides a high-performance version of base R’s data.frame with syntax and feature enhancements for ease of use, convenience and programming speed.” The dplyr library (Wickham, François, et al. 2023) implements “a grammar of data manipulation” and forms part of the Tidyverse, which is “an opinionated collection of R packages designed for data science.”

While standard Python does not include a data frame library, there are several third-party data frame libraries used for data analysis in Python. By far the most established data frame library is pandas, which has been the de facto standard data frame library in Python for more than a decade. Its broad adoption means that much existing Python code, documentation, and discussion of data analysis assumes some familiarity with pandas.

In recent years, other data frame libraries have emerged to address perceived limitations of pandas or to support special use cases. For example, PySpark provides a data frame interface designed for distributed computing across clusters, making it useful for very large data sets. Ibis offers a higher-level interface for working with data stored in databases. Recently Polars has attracted attention for its speed, efficiency, and expressive syntax.

In this book, I focus on Polars because its syntax makes data-manipulation code relatively easy to read while also scaling well to larger data sets. In addition, as I discuss below, Polars provides facility for adding namespaces that allows me to use a more fluent methods-based approach to plotting.

Data visualization libraries

R

Base graphics in R were written by Ross Ihaka drawing on experience implementing the graphics driver of S, a predecessor of R, and Chambers et al. (1983). Base graphics functions are generally fast, but have limited scope and functionality.

The development of “grid” graphics, a much richer system of graphical primitives, started in 2000 The lattice package, developed by Deepayan Sarkar (Sarkar 2008}), uses grid graphics to implement the trellis graphics system of Cleveland (1993) and is a considerable improvement over base graphics. You can easily produce conditioned plots and some plotting details are taken care of automatically. However, lattice graphics lack a formal model, which extending them more difficult.

The ggplot2, started in 2005, is an attempt to take the good things about base and lattice graphics and improve on them with a strong underlying model which supports the production of any kind of statistical graphic, based on the principles outlined above. The solid underlying model of ggplot2 makes it easy to describe a wide range of graphics with a compact syntax, and independent components make extension easy. Like lattice, ggplot2 uses grid to draw the graphics, which means you can exercise much low-level control over the appearance of the plot.

Python

Barrett, Tyson, Matt Dowle, Arun Srinivasan, et al. 2026. Data.table: Extension of ‘Data.frame‘. https://doi.org/10.32614/CRAN.package.data.table.

Chambers, John, William Cleveland, Beat Kleiner, and Paul Tukey. 1983. Graphical Methods for Data Analysis. Wadsworth. https://doi.org/10.1201/9781351072304.

Cleveland, William. 1993. Visualizing Data. Hobart Press.

Healy, Kieran. 2026. Data Visualization: A Practical Introduction. 2nd ed. Princeton University Press. https://socviz.co/.

McKinney, Wes. 2022. Python for Data Analysis. 3rd ed. O’Reilly Media.

Sarkar, Deepayan. 2008. Lattice: Multivariate Data Visualization with r. Springer. https://doi.org/10.1007/978-0-387-75969-2.

Wickham, Hadley, Mine Çetinkaya-Rundel, and Garrett Grolemund. 2023. R for Data Science. 2nd ed. O’Reilly Media.

Wickham, Hadley, Romain François, Lionel Henry, Kirill Müller, and Davis Vaughan. 2023. Dplyr: A Grammar of Data Manipulation. https://doi.org/10.32614/CRAN.package.dplyr.

---

1. The goal of the pandas package was to provide just such a library for Python, which had previously lacked one. Since the release of pandas, Python has gone on to become the most popular language for data science.

2. The dplyr documentation discusses a version of this list.