Matplotlib + Pandas =如何看标签？

Question

我正在尝试显示带有长标签的数据框。 当我希望它显示标签时，情节主要由图表占据。 我有：

new_labels = []
for i, index in enumerate(df.index):
    new_label = "%s (%.2f)"%(index,df.performance[i])
    new_labels.append(new_label)

fig , axes = plt.subplots(1,1)
df.sort_values(col_name).plot(kind='barh', ax=axes)
axes.xaxis.set_ticklabels(new_labels)
axes.xaxis.set_major_locator(ticker.MultipleLocator(1))

这给了我：

如您所见，标签不会显示。 确实有价值观：

new_labels
['PLS regression\n\n    PLSRe (0.12)',
 'Regression based on k-nea (0.44)',
 'The Gaussian Process mode (0.46)',
 'Orthogonal Matching Pursu (0.52)',
 'An extremely randomized t (0.54)',
 'RANSAC (RANdom SAmple Con (0.56)',
 'Elastic Net model with it (0.66)',
 'Kernel ridge regression. (0.67)',
 'Cross-validated Orthogona (0.67)',
 'Linear Model trained with (0.68)',
 'Linear regression with co (0.68)',
 'Theil-Sen Estimator (0.68)',
 'Lasso linear model with i (0.69)',
 'Bayesian ridge regression (0.70)'...

如何为标签提供更多空间，并使用更短的条形图？

Answer 1

试试这个：

df.sort_values(col_name).plot(x=new_labels, kind='barh', ax=axes)
#  NOTE:                      ^^^^^^^^^^^^

Answer 2

您可以在绘图的左侧添加更多空间，以便标签有机会完全显示。这可以使用

完成

fig.subplots_adjust(left=0.5)

其中0.5表示数字宽度的50％。

完整示例：

import pandas as pd
import matplotlib.pyplot as plt

labels = ['PLS regression PLSRe (0.12)',
 'Regression based on k-nea (0.44)',
 'The Gaussian Process mode (0.46)',
 'Orthogonal Matching Pursu (0.52)',
 'An extremely randomized t (0.54)',
 'RANSAC (RANdom SAmple Con (0.56)',
 'Elastic Net model with it (0.66)',
 'Kernel ridge regression. (0.67)',
 'Cross-validated Orthogona (0.67)',
 'Linear Model trained with (0.68)',
 'Linear regression with co (0.68)',
 'Theil-Sen Estimator (0.68)',
 'Lasso linear model with i (0.69)',
 'Bayesian ridge regression (0.70)']

values = [0.12, .44,.46,.52,.54,.56,.66,.67,.67,.68,.68,.68,.69,.7]

df = pd.DataFrame(values, index=labels)

fig , axes = plt.subplots(1,1)
fig.subplots_adjust(left=0.5)
df.plot(kind='barh', ax=axes)

plt.show()