Scratch Coding and Game Development for Tianmu Expat Kids in Taipei

1. Why Heatmaps for Correlation

A correlation matrix is numeric; a heatmap turns those values into a color grid so strong, weak, and inverse pairs are easier to scan as feature count grows.

• Use only numeric columns for .corr()
• Diagonal is always 1.0 (self-correlation)
• Off-diagonal values lie between −1 and +1

2. Build the Matrix

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np

df = pd.read_csv("healthcare_patient_analytics_seaborn.csv")
numeric_df = df.select_dtypes(include=["number"])
corr_matrix = numeric_df.corr()
print(corr_matrix)

Magnitude shows strength; sign shows direction (same way vs opposite).

3. Full Heatmap

plt.figure(figsize=(8, 6))
sns.heatmap(
    corr_matrix,
    annot=True,
    cmap="coolwarm"
)
plt.title("Healthcare Feature Correlation Heatmap")
plt.show()

annot=True prints coefficients in cells. cmap="coolwarm" maps near-zero/negative toward blue and values near 1 toward red.

4. Mask the Redundant Half

The matrix is symmetric; the upper triangle mirrors the lower. Hide duplicates and the diagonal with np.triu:

mask = np.triu(np.ones_like(corr_matrix, dtype=bool))

plt.figure(figsize=(8, 6))
sns.heatmap(
    corr_matrix,
    mask=mask,
    annot=True,
    cmap="coolwarm"
)
plt.title("Half Correlation Heatmap (Lower Triangle Only)")
plt.show()