Altair - How to create a grouped bar Chart without manipulating the DataFrame

Question

I noticed that one can create grouped bar charts in Altair, by evaluating the column of a DataFrame. My issue is that my data doesn't have each group as a value of an specific column.

So, is it possible to use the column name as a group name in Altair (without evaluating the column) or modifying the DataFrame?

Below is the DataFrame I have and the grouped bar chart I need:

Answer 1

This is an example of wide-form data (See Long-form vs. Wide-form Data ). To transform it to Long-form data without modifying the dataframe, you can use the Fold Transform .

Once you've done this, you can follow the Grouped Bar Chart Example to make your chart. It might look something like this:

import pandas as pd
import altair as alt

df = pd.DataFrame({
    "Job Stat": ['INV', "WRK", "CMP", "JRB"],
    "Revenue": [100, 200, 300, 400],
    "Total Income": [150, 250, 350, 450]
})

(
  alt.Chart(df)
    .transform_fold(["Revenue", "Total Income"], as_=["key", "value"])
    .mark_bar()
    .encode(
        x="key:N",
        y="value:Q",
        color="key:N",
        column="Job Stat",
    )
)

To make it closer to your example chart, you can tweak some of the label & legend settings:

(
  alt.Chart(df)
    .transform_fold(["Revenue", "Total Income"], as_=["key", "value"])
    .mark_bar()
    .encode(
        alt.X('key:N', axis=None),
        alt.Y("value:Q"),
        alt.Color("key:N", legend=alt.Legend(title=None, orient='bottom')),
        alt.Column("Job Stat",
          sort=['INV', "WRK", "CMP", "JRB"],
          header=alt.Header(labelOrient="bottom", title=None)
        )
    )
)