Quarto

Quarto weaves together narrative text and code to produce elegantly formatted output. Most interestingly, Quarto executes code written in different languages. The following gives an illustration of cross-language coding in Quarto (here using R and Python - other languages, e.g. Julia, are supported as well).

Cross-language coding

R

R code chunks are written and executed just as we would do it in R Markdown. So, for users of R Markdown this should be very familiar.

if (!require("pacman")) install.packages("pacman")
pacman::p_load(tidyverse, palmerpenguins, reticulate, viridis)

data(penguins, package = "palmerpenguins")
penguins <- penguins[complete.cases(penguins),]

ggplot(penguins, 
       aes(x = flipper_length_mm, y = bill_length_mm)) +
  geom_point(aes(color = species, shape = species)) +
  scale_color_viridis(discrete = TRUE) +
  labs(
    title = "Flipper and bill length",
    subtitle = "Dimensions for penguins at Palmer Station LTER",
    x = "Flipper length (mm)", y = "Bill length (mm)",
    color = "Penguin species", shape = "Penguin species"
  ) +
  theme_light()

Python

The beauty comes into play when we add Python chunks of code.

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import classification_report

penguins = r.penguins # here the R tibble turns into a pandas DataFrame
penguins.head(3)

species	island	bill_length_mm	...	body_mass_g	sex	year
Adelie	Torgersen	39.1	...	3750	male	2007
Adelie	Torgersen	39.5	...	3800	female	2007
Adelie	Torgersen	40.3	...	3250	female	2007

We simply switched from R to Python working on the penguins data set.

Now, let us train a simple classifier using the scikit-learn module in Python.

y = penguins['species']
X = penguins['bill_length_mm','bill_depth_mm','flipper_length_mm','body_mass_g']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=333)

classifier = RandomForestClassifier().fit(X_train, y_train)
y_test_pred = classifier.predict(X_test)

print(classification_report(y_test, y_test_pred))

	Precision	Recall	F1-Score	Support
Adelie	1.00	0.95	0.98	22
Chinstrap	1.00	1.00	1.00	15
Gentoo	0.97	1.00	0.98	30
Average/Total	0.99	0.99	0.99	67

Conclusion

Great, with Quarto it is possible to write reports, presentations, … using different languages. Here, we have made use of R's tidyverse for data preprocessing and visualization and Python's Machine Learning frameworks (scikit-learn in this example). So, Quarto really enables you to pick the best of both worlds.