seaborn 点图中的变量标记大小

Question

我想根据代表样本大小的 numpy 数组来改变我的 seaborn 点图中的标记大小。

markersize_array = np.array([[10, 5, 6]])

将

markersize_array

提供给

markersize

关键字会产生错误：

TypeError: only length-1 arrays can be converted to Python scalars

这是一些示例代码：

penguins = sns.load_dataset("penguins")

fig, ax = plt.subplots(1, 1, figsize=(2, 2))

markersize_array = np.array([[10, 5, 6]])

sns.pointplot(
    data=penguins,
    x="body_mass_g",
    y="island",
    ax=ax,
    estimator=np.mean,
    errorbar=('ci', 95),
    capsize=0,
    color="k",
    log_scale=False,
    linestyle="none",
    marker="s",
    markersize=np.array([[5]]),
    # markersize=markersize_array,  # <- I need this
)

有办法做到这一点吗？

编辑

使用seaborn对象也不起作用：

import seaborn.objects as so

tips = sns.load_dataset("tips")

p1 = so.Plot(tips, "total_bill", "tip")
p1.add(so.Dot())
(
    so.Plot(tips, x="total_bill", y="day")
    .add(so.Dot(pointsize=np.array(10)), so.Agg())
    # .add(so.Dot(pointsize=np.array([[10, 10, 10, 10]])), so.Agg())
    .add(so.Range(), so.Est(errorbar=("se", 2)))
)

Answer 1

ChatGPT 为我提供了一种使用 matplotlib 的解决方法，该方法需要很多行代码但达到了预期的结果：

import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np
import scipy.stats as stats

# Load the example dataset
penguins = sns.load_dataset("penguins")

# Example data to simulate markersize array
markersize_array = np.array([10, 5, 6])

# Calculate the means and sample sizes
grouped_data = penguins.groupby("island")["body_mass_g"].agg(["mean", "count"])

ci_bounds = penguins.groupby("island")["body_mass_g"].apply(
    lambda x: sns.utils.ci(sns.algorithms.bootstrap(x, func=np.mean), 95)
)

# Create the plot
fig, ax = plt.subplots(figsize=(5, 5))

# Iterate through the islands and plot each point individually with the corresponding marker size
for i, island in enumerate(grouped_data.index):
    # Extract the mean, CI lower, and upper bounds
    mean_val = grouped_data.loc[island, "mean"]
    ci_lower, ci_upper = ci_bounds[island]

    # Plot the point
    ax.errorbar(
        mean_val,  # x value (mean)
        island,  # y value (island)
        xerr=[[mean_val - ci_lower], [ci_upper - mean_val]],  # CI as error bars
        fmt='o',  # marker style
        color='k',  # color of the point
        markersize=markersize_array[i],  # marker size from array
        capsize=5,  # size of the caps on the error bars
    )

# Final plot adjustments
ax.set_xlabel("Body Mass (g)")
ax.set_title("Body Mass by Island with Varying Marker Sizes")
sns.despine()

plt.tight_layout()
plt.show()