Heatmap

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

# === USER-EDITABLE PARAMETERS ===
# DataFrame to use (replace with your own if needed)
df = None  # Change: Set to your DataFrame if not generating example data

# Example dataset parameters (used if df is None)
n_samples = 200  # Change: Number of rows in the example dataset
n_features = 6   # Change: Number of variables in the example dataset
random_seed = 42  # Change: Seed for reproducibility

# Plotting parameters
figsize = (8, 6)          # Change: Figure size (width, height) in inches
title = "Correlation Matrix Heatmap: Strong Positive & Negative Relationships"  # Change: Insight-driven title
cmap = "RdBu"             # Change: Diverging colormap (e.g., 'RdBu', 'coolwarm')
annot = True              # Change: Show correlation coefficients
fmt = ".2f"               # Change: Format for annotation numbers
center = 0                # Change: Center value for diverging colormap
linewidths = 0.5          # Change: Width of lines between cells
cbar_shrink = 0.8         # Change: Shrink factor for colorbar
title_fontsize = 18       # Change: Font size for title
label_fontsize = 16       # Change: Font size for axis labels and ticks

# === END-OF-CODE ===

# === Data preparation ===
# If no DataFrame provided, generate a synthetic example dataset
if df is None:
    np.random.seed(random_seed)
    # Create random data with some correlations
    rng = np.random.default_rng(random_seed)
    # Generate base random matrix
    base = rng.standard_normal(size=(n_samples, n_features))
    # Introduce correlations by linear combinations
    df = pd.DataFrame(base, columns=[f"Var{i+1}" for i in range(n_features)])
    # Add a positive correlation between Var1 and Var2
    df["Var2"] = df["Var1"] * 0.8 + rng.standard_normal(n_samples) * 0.2
    # Add a negative correlation between Var3 and Var4
    df["Var4"] = -df["Var3"] * 0.6 + rng.standard_normal(n_samples) * 0.4

# Compute correlation matrix
try:
    corr_matrix = df.corr()
except Exception as e:
    print(f"Error computing correlation matrix: {e}")
    corr_matrix = pd.DataFrame()

# Print correlation matrix for reference
print("Correlation matrix:")
print(corr_matrix)

# Print maximum absolute correlation (excluding diagonal)
if not corr_matrix.empty:
    max_corr = corr_matrix.where(~np.eye(corr_matrix.shape[0], dtype=bool)).abs().max().max()
    print(f"Maximum absolute correlation (excluding diagonal): {max_corr:.2f}")

# === Create plot ===
fig, ax = plt.subplots(figsize=figsize)

# Plot heatmap
sns.heatmap(
    corr_matrix,
    cmap=cmap,
    center=center,
    annot=annot,
    fmt=fmt,
    square=True,
    linewidths=linewidths,
    cbar_kws={"shrink": cbar_shrink},
    ax=ax
)

# Set title and adjust font sizes
ax.set_title(title, fontsize=title_fontsize, pad=12)
ax.tick_params(labelsize=label_fontsize)

# Adjust layout to avoid clipping
plt.subplots_adjust(top=0.90, bottom=0.12, left=0.12, right=0.95)
plt.tight_layout()

# Assign figure to variable for consistency
fig = ax.figure

# Show plot
plt.show()
# END-OF-CODE

Strong Positive Correlations (>0.7):
  Revenue - Profit: 0.85
  Revenue - Customers: 0.88
  Profit - ROI: 0.92
  Revenue - Marketing: 0.72
  Customers - AOV: 0.72

Strong Negative Correlations (<-0.4):
  Revenue - Expenses: -0.45
  Profit - Expenses: -0.78
  Expenses - ROI: -0.65