Setting order of categorical columns in Altair histogram

Question

Trying to swap order of columns in histogram.

data = {'Age group': {0: '0 - 4', 1: '12 - 16', 2: '5 - 11', 3: 'not specified'},
 'Count': {0: 81.0, 1: 86.0, 2: 175.0, 3: 0.0}}

dp = pd.DataFrame(data)

alt.Chart(dp).mark_bar().encode(
    x=alt.X('Age group:N', title='Age group'),
    y='Count:Q',
    tooltip=['Age group', 'Count']
).properties(title='Number of children in voucher', width=400)

Just want to swap the order and can't find a way of doing it.

Answer 1

The Alt.X schema has a sort parameter that accepts an array specifying the field values in preferred order :

X_ORDER = ["0 - 4", "5 - 11", "12 - 16", "not specified"] # >> set an order

alt.Chart(dp).mark_bar().encode(
    x=alt.X(
        "Age group:N",
        title="Age group",
        sort=X_ORDER, # >> add this line
    ),
    y="Count:Q",
    tooltip=["Age group", "Count"],
).properties(title="Number of children in voucher", width=400)

Output :

Answer 2

In addition to setting the category order directly in altair via sort, altair is also able to infer the order from the pandas dataframe when you work with categorical data (which can be preferable for other pandas operations as well):

import pandas as pd
import altair as alt


data = pd.DataFrame({
    'Age group': {0: '0 - 4', 1: '12 - 16', 2: '5 - 11', 3: 'not specified'},
    'Count': {0: 81.0, 1: 86.0, 2: 175.0, 3: 0.0}
})

data['Age group'] = pd.Categorical(
    data['Age group'],
    ['0 - 4', '5 - 11', '12 - 16', 'not specified'],
    ordered=True
)

alt.Chart(data, title='Number of children in voucher', width=200).mark_bar().encode(
    x='Age group',
    y='Count:Q',
    tooltip=['Age group', 'Count']
)