Seaborn Regression Plots

Regression Plots in Seaborn

Understanding relationships between variables is crucial for data science.

regplot – simple linear regression

sns.regplot(x='hours', y='score', data=tips, scatter_kws={'alpha':0.6}, line_kws={'color':'red'})
plt.title('Regression of Score on Study Hours')
plt.show()

lmplot – regression with facets

sns.lmplot(x='hours', y='score', hue='sex', data=tips, col='day', height=4, aspect=1)
plt.show()

Residual plot

sns.residplot(x='hours', y='score', data=tips, lowess=True, color='green')
plt.title('Residual Plot')
plt.show()

Polynomial regression (degree=2)

sns.lmplot(x='hours', y='score', data=tips, order=2, ci=None, line_kws={'color':'purple'})
plt.show()

Interpreting the plots

• Slope indicates direction/magnitude of relationship.
• Confidence interval (shaded area) shows uncertainty.
• Residuals help assess model fit; random scatter suggests a good fit.