python - 重用基础图而无需重新绘制

Question

我有一个大型数据集，想要将整个数据集绘制为背景，然后通过子集化并在背景顶部绘制来突出显示其中的过滤特征。我通过每次重新绘制背景来完成此工作，但这非常耗时，因为我基于此渲染了大约 40 个图。

我遇到的问题是我似乎无法让背景数据(第一个散点图)保持在原位。通过复制图形或尝试复制轴。

完整功能代码示例:

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt


df = pd.DataFrame(
    {
        "x": np.random.normal(size=100),
        "y": np.random.rand(100),
        "thing_1": np.concatenate((np.ones(50), np.zeros(50))),
        "thing_2": np.concatenate((np.zeros(50), np.ones(50)))}
)

fig, ax = plt.subplots(figsize=(12, 8))


# This works but replots the background data each time (costly with the large datasets)
for thing in ['thing_1', 'thing_2']:

    ax.clear()
    # background data cloud  Reuse instead of plotting
    ax.scatter(df.x, df.y, c='grey', alpha=0.5, s=30)

    # subset to highlight
    ind = df[thing] == 1
    ax.scatter(df.loc[ind, 'x'], df.loc[ind, 'y'], c='red', alpha=1, s=15)

    plt.savefig('{}_filter.png'.format(thing))

我目前优化代码的最佳尝试:

# Want to do something like this (only plot background data once and copy the axis or figure)
fig_background, ax_background = plt.subplots(figsize=(12, 8))
ax_background.scatter(df.x, df.y, c='grey', alpha=0.5, s=30)

for thing in ['thing_1', 'thing_2']:
    fig_filter = fig_background

    axs = fig_filter.get_axes()

    # subset to highlight
    ind = df[thing] == 1
    axs[0].scatter(df.loc[ind, 'x'], df.loc[ind, 'y'], c='red', alpha=1, s=15)

    plt.savefig('{}_filter.png'.format(thing))

    plt.cla()

Answer 1

您可以在绘制新循环步骤之前删除每个循环步骤中的散点。

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt


df = pd.DataFrame(
    {
        "x": np.random.normal(size=100),
        "y": np.random.rand(100),
        "thing_1": np.concatenate((np.ones(50), np.zeros(50))),
        "thing_2": np.concatenate((np.zeros(50), np.ones(50)))}
)

fig, ax = plt.subplots(figsize=(12, 8))
# background data cloud
ax.scatter(df.x, df.y, c='grey', alpha=0.5, s=30)

scatter = None

for thing in ['thing_1', 'thing_2']:

    if scatter is not None:
        scatter.remove()

    # subset to highlight
    ind = df[thing] == 1
    scatter = ax.scatter(df.loc[ind, 'x'], df.loc[ind, 'y'], c='red', 
    alpha=1, s=15)

    plt.savefig('{}_filter.png'.format(thing))