objectif

Explorer corrélations via pairplot et matrice de corr.

code minimal

import seaborn as sns

df = sns.load_dataset("iris")
g = sns.pairplot(df, hue="species", corner=True)
print(hasattr(g, "fig"))

utilisation

import seaborn as sns
import numpy as np

df = sns.load_dataset("iris").drop(columns=["species"])
corr = df.corr(numeric_only=True)
sns.heatmap(corr, annot=False)
print(corr.shape)

variante(s) utile(s)

import seaborn as sns

df = sns.load_dataset("tips")
sns.pairplot(df.select_dtypes("number"))
print("ok")

notes

• Attention aux volumes: sous-échantillonner si nécessaire.