python seaborn heatmap可视化相关性矩阵

方法

import pandas as pd
import numpy as np
import seaborn as sns
df = pd.DataFrame(np.random.randn(50).reshape(10,5))
corr = df.corr()
sns.heatmap(corr, cmap='Blues', annot=True)

将矩阵型简化为对角矩阵型：

mask = np.zeros_like(corr)
mask[np.tril_indices_from(mask)] = True
sns.heatmap(corr, cmap='Blues', annot=True, mask=mask.T)