Altair：用连续比例着色的散点图的回归

Question

我正在尝试使用 Altair 显示散点图，其中标记颜色由非分类特征（连续）给出，并在其上添加回归线。不过，我对 Altair 还很陌生。

第一，here's a sample of the data I'm inputting into Altair.

So far I can create either the colouring without a regression:

问题是回归线没有显示出来。这是通过以下代码实现的：

chart = alt.Chart(df).mark_circle(size=60).encode(
   x='Lambda',
   y='ACR',
   color='Consensus',
   tooltip=["Tolerance Prevalence", "Consensus", "ACR", "Lambda"]
)

chart.interactive() + chart.transform_regression(
   'Lambda', 'ACR', method="quad"
).mark_line(color="red")

Or a regression without the desired colour scale. 我可以通过简单地删除第一条指令中的“color='consensus'”行来做到这一点。

我试过改变回归方法，甚至不同的特征组合都无济于事。我可以使用任何参数或 Altair 函数轻松解决此问题吗？

提前致谢！

编辑 1

完整代码：

import altair as alt
import pandas as pd
import numpy as np

df = read_csv("processed.txt")

chart = alt.Chart(df).mark_circle(size=60).encode(
    x='Lambda',
    y='ACR',
    color='Consensus',
    tooltip=["Tolerance Prevalence", "Consensus", "ACR", "Lambda"]
)

chart.interactive() + chart.transform_regression(
    'Lambda', 'ACR', method="quad").mark_line(color="red")

chart

Full Data (processed.txt)

Answer 1

您可以定义一个没有着色的基本图表，然后从中构建散点图和回归线：

import altair as alt
import pandas as pd
import numpy as np


df = pd.read_csv("Downloads/processed.txt", sep=' ')

base = alt.Chart(df).mark_circle(size=60).encode(
    x='Lambda',
    y='ACR',
)

scatter = base.encode(
    color='Consensus',
    tooltip=["Tolerance Prevalence", "Consensus", "ACR", "Lambda"]
)

line = base.mark_line(color="red").transform_regression(
    'Lambda', 'ACR', method="quad")

scatter + line

Altair：用连续比例着色的散点图的回归

Altair: Regression over a scatter plot coloured with a continuous scale

python

regression

data-visualization

scatter-plot

altair