带抖动的散点图

在此图表中，我们将 movies 数据集中的 Major_Genre 列编码到 y 通道。在该数据的默认表示中，由于重叠太多，很难衡量不同值出现的相对频率。为了解决这个问题，我们使用 yOffset 通道来加入随机偏移（抖动）。该示例显示了两次，左侧使用正态分布的抖动，右侧使用均匀分布的抖动。

方法语法

import altair as alt
from vega_datasets import data

source = data.movies.url

gaussian_jitter = alt.Chart(source, title='Normally distributed jitter').mark_circle(size=8).encode(
    y="Major_Genre:N",
    x="IMDB_Rating:Q",
    yOffset="jitter:Q",
    color=alt.Color('Major_Genre:N').legend(None)
).transform_calculate(
    # Generate Gaussian jitter with a Box-Muller transform
    jitter="sqrt(-2*log(random()))*cos(2*PI*random())"
)

uniform_jitter = gaussian_jitter.transform_calculate(
    # Generate uniform jitter
    jitter='random()'
).encode(
    alt.Y('Major_Genre:N').axis(None)
).properties(
    title='Uniformly distributed jitter'
)

(gaussian_jitter | uniform_jitter).resolve_scale(yOffset='independent')

属性语法

import altair as alt
from vega_datasets import data

source = data.movies.url

gaussian_jitter = alt.Chart(source, title='Normally distributed jitter').mark_circle(size=8).encode(
    y="Major_Genre:N",
    x="IMDB_Rating:Q",
    yOffset="jitter:Q",
    color=alt.Color('Major_Genre:N', legend=None)
).transform_calculate(
    # Generate Gaussian jitter with a Box-Muller transform
    jitter="sqrt(-2*log(random()))*cos(2*PI*random())"
)

uniform_jitter = gaussian_jitter.transform_calculate(
    # Generate uniform jitter
    jitter='random()'
).encode(
    y=alt.Y('Major_Genre:N', axis=None)
).properties(
    title='Uniformly distributed jitter'
)

(gaussian_jitter | uniform_jitter).resolve_scale(yOffset='independent')