Seaborn：在直方图上用误差条覆盖箱形图或均值 [英] Seaborn: Overlaying a box plot or mean with error bars on a histogram

问题描述

我正在以非常标准的方式在Seaborn中创建数据的直方图，即：

I am creating a histogram in Seaborn of my data in a pretty standard way, ie:

rc = {'font.size': 32, 'axes.labelsize': 28.5, 'legend.fontsize': 32.0, 
    'axes.titlesize': 32, 'xtick.labelsize': 31, 'ytick.labelsize': 12}
sns.set(style="ticks", color_codes=True, rc = rc)
plt.figure(figsize=(25,20),dpi=300)

ax = sns.distplot(synData['SYNERGY_SCORE'])
print (np.mean(synData['SYNERGY_SCORE']), np.std(synData['SYNERGY_SCORE']))
# ax = sns.boxplot(synData['SYNERGY_SCORE'], orient = 'h') 

ax.set(xlabel = 'Synergy Score', ylabel = 'Frequency', title = 'Aggregate Synergy Score Distribution')

这将产生以下输出：

我还想在同一图中可视化此数据集的平均值+标准差在x轴上（或x轴正上方）有一个均值点和带缺口的误差棒显示标准偏差。另一种选择是包围x轴的箱线图。我尝试仅添加注释掉的行（sns.boxplot（）），但它看起来非常丑陋，根本不符合我的要求。有建议吗？

I also want to visualize the mean + standard deviation of this dataset on the same plot, ideally by having a point for the mean on the x-axis (or right above the x-axis) and notched error bars showing the standard deviation. Another option is a boxplot hugging the x-axis. I tried just adding the line which is commented out (sns.boxplot()), but it looks super ugly and not at all what I'm looking for. Any suggestions?

推荐答案

箱线图是在分类轴上绘制的，不会与直方图的密度轴很好地共存，但可以使用x轴双轴图来完成：

The boxplot is drawn on a categorical axis and won't coexist nicely with the density axis of the histogram, but it's possible to do it with a twin x axis plot:

import numpy as np
import seaborn as sns

x = np.random.randn(300)
ax = sns.distplot(x)
ax2 = ax.twinx()
sns.boxplot(x=x, ax=ax2)
ax2.set(ylim=(-.5, 10))

这篇关于Seaborn：在直方图上用误差条覆盖箱形图或均值的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！