Sample Transform

The sample transform is one of the simpler of all Altair’s data transforms; it takes a single parameter sample which specified a number of rows to randomly choose from the dataset. The resulting chart will be created using only this random subset of the data.

For example, here we chart the full cars dataset alongside a sample of 100 rows:

import altair as alt
from vega_datasets import data

source =

chart = alt.Chart(source).mark_point().encode(

chart | chart.transform_sample(100)

Transform Options

The transform_sample() method is built on the SampleTransform class, which has the following options:

Property Type Description
sample number

The maximum number of data objects to include in the sample.

Default value: 1000