🔥 Heatmap

Step 1: Set index to country names

df = df.set_index('Countries')

Step 2: Initial heatmap (includes all columns, including life expectancy)

sns.heatmap(df)

Step 3: Drop Life expectancy column for clearer numerical focus

rates = df.drop('Life expectancy', axis=1)

Step 4: Adjust visual details

sns.heatmap(rates)  # Default
sns.heatmap(rates, linewidth=0.5)  # Add cell borders
sns.heatmap(rates, linewidth=0.5, annot=True)  # Show numbers

Step 5: Customize with color palettes

sns.heatmap(rates, linewidth=0.5, annot=True, cmap='viridis')

Step 6: Center colorbar on a specific value

sns.heatmap(rates, linewidth=0.5, annot=True, cmap='viridis', center=40)  # Center at 40
sns.heatmap(rates, linewidth=0.5, annot=True, cmap='viridis', center=1)   # Center at 1

🔍 center helps highlight deviation from a key value (e.g., population growth rate = 1) but don’t recommend using center method, default is fine

---

🧬 Clustermap

Use clustermap to group similar rows/columns using hierarchical clustering.

sns.clustermap(rates)

Turn off column clustering (for fixed column order):

sns.clustermap(rates, col_cluster=False)

Adjust figure size and colorbar position:

sns.clustermap(
    rates,
    col_cluster=False,
    figsize=(12, 8),
    cbar_pos=(-0.1, .2, .03, .4)  # [left, bottom, width, height]
)

Optional: Clean up row index name

rates.index.set_names('', inplace=True)

---

✅ Key Takeaways

• heatmap is good for visualizing raw matrix data.

• clustermap is ideal for discovering patterns/groupings.

• Use annot=True to display actual values.

• Use center to customize midpoint in colormap (useful for diverging palettes).

• Always clean/preprocess your DataFrame before plotting.