Correlation matrix plot with coefficients on one side, scatterplots on another, and distributions on diagonal

Question

I love this correlation matrix from the PerformanceAnalytics R package\'s chart.Correlation function:

How can I create this in Python? The corr

Answer 1

The cor_matrix function below does this, plus adds a bivariate kernel density plot. Thanks to @karl-anka's comment for getting me started.

import matplotlib.pyplot as plt
import seaborn as sns
from scipy import stats

sns.set(style='white')
iris = sns.load_dataset('iris')

def corrfunc(x, y, **kws):
  r, p = stats.pearsonr(x, y)
  p_stars = ''
  if p <= 0.05:
    p_stars = '*'
  if p <= 0.01:
    p_stars = '**'
  if p <= 0.001:
    p_stars = '***'
  ax = plt.gca()
  ax.annotate('r = {:.2f} '.format(r) + p_stars,
              xy=(0.05, 0.9), xycoords=ax.transAxes)

def annotate_colname(x, **kws):
  ax = plt.gca()
  ax.annotate(x.name, xy=(0.05, 0.9), xycoords=ax.transAxes,
              fontweight='bold')

def cor_matrix(df):
  g = sns.PairGrid(df, palette=['red'])
  # Use normal regplot as `lowess=True` doesn't provide CIs.
  g.map_upper(sns.regplot, scatter_kws={'s':10})
  g.map_diag(sns.distplot)
  g.map_diag(annotate_colname)
  g.map_lower(sns.kdeplot, cmap='Blues_d')
  g.map_lower(corrfunc)
  # Remove axis labels, as they're in the diagonals.
  for ax in g.axes.flatten():
    ax.set_ylabel('')
    ax.set_xlabel('')
  return g

cor_matrix(iris)