Altair: Using Altair with RMarkdown

[zacharylawrence]

Hi - is it currently possible to use Altair when rendering python chunks in RMarkdown?

For example, I have the RMarkdown file:

````rmarkdown

title: "Test"

output: pdf_document

```{python}
import pandas as pd
import altair as alt

df = pd.DataFrame({'col1': [1, 2], 'col2': [3, 4]})

```{python}
df.plot()

```{python}
alt.Chart(df).mark_point()

````

and render it with:
```bash
Rscript -e "rmarkdown::render('test.Rmd')"

The 2nd code chunk display a matplotlib graphic in the output pdf, but the altair chart is just displayed as ## alt.Chart(...).

I suspect this is due to Altair "rendering" a vega-lite schema or JavaScript? However since I'm rendering to a PDF, I would like to simply embed a static image. I know Altair can do this with alt.Chart(df).mark_point().save('test.png'), but I would like to embed the resulting image into the rendered RMarkdown PDF instead of to a separate file.

Do you have any advice on if this is possible or next steps?

As a followup question, would rending the RMarkdown file to HTML instead of PDF (allowing for Altair interactivity) be possible?

Thanks!

[jakevdp]

I'm not familiar enough with rmarkdown to know how to inject the JS necessary to render charts, but you may be able to find some clues in this issue: https://github.com/altair-viz/altair/issues/1141

The altair CRAN package may also be useful: https://cran.rstudio.com/web/packages/altair/index.html

[firasm]

With some help from the folks at vegawidget, I was able to get Altair working in an Rmd html and pdf, below is the Rmarkdown code:

---
title: "Altair-reticulate"
output: html_document
---

```{r load_packages}
library(vegawidget) # Need to install this using: install.packages('vegawidget')
library(reticulate)
use_python('/usr/local/bin/python3') # you may not need this, comment out if so

```{python Altair}
import altair as alt
import pandas as pd
source = pd.DataFrame({
'a': ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I'],
'b': [28, 55, 43, 91, 81, 53, 19, 87, 52]
})
chart1 = alt.Chart(source).mark_bar().encode(
alt.X('a:N',title='X-Axis label'),
alt.Y('b:Q',title='Y-axis label')).to_json()

```{r display1}
as_vegaspec(py$chart1)

```{python Altair4}
import altair as alt
import pandas as pd
source = pd.DataFrame({
'a': ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I'],
'b': [28, 55, 43, 91, 81, 53, 19, 87, 52]
})
chart2 = alt.Chart(source).mark_bar().encode(
alt.X('a:N',title='X-Axis label'),
alt.Y('b:Q',title='Y-axis label')
).properties(title={
"text": ["Altair now has multi-line titles!!", "Second line of title"],
"subtitle": ["And subtitles! SO COOL", "Even cooler second line wow dang, but you need to update to v4 of Altair"],
"color": "red",
"subtitleColor": "green"
}).to_json()

```{r display2}
as_vegaspec(py$chart2)

h/t to @ijlyttle from this issue

[zacharylawrence]

Thanks for your thoughts on this! I haven't gotten the chance to try this but will go ahead and close the issue (and will re-open if I can't get it working).

Again - thanks!

[joelostblom]

Thanks! This works for me for displaying inside RStudio, but not in knitted HTML, which might be a version issue. If you would like RStudio to support Altair/Vega plots natively in RMarkdown, please give your thumbs up on this issue https://github.com/rstudio/rstudio/issues/4259.