confusion_matrix

Plot a confusion matrix heatmap with optional normalization and annotations.

---

📥 Arguments

Name	Type	Required	Description
cm	pd.DataFrame	✅	Confusion matrix as a pandas DataFrame (shape N×N).
labels	List[str]	❌	List of class labels, length N, corresponding to matrix indices. If None, use DataFrame index/columns.
figsize	tuple	❌	Figure size, e.g., (8, 6).
cmap	str	❌	Matplotlib colormap name, e.g., ‘Blues’.
normalize	bool	❌	If True, normalize rows to sum to 1 before plotting.
title	str	❌	Optional title for the plot.
fmt	str	❌	Format string for annotations, e.g., ‘d’ or ‘.2f’.
cbar	bool	❌	Whether to display the colorbar.
save	str	❌	Base filename to save PNG and PDF outputs.
fontsize	float	❌	Base font size for annotations, ticks, and title.

---

📦 Example Output

Click to show example code

import pandas as pd
import matplotlib.pyplot as plt
from swizz import plot

# 1) Define a toy confusion matrix as a DataFrame
data = [
    [50,  2,  3],
    [ 5, 45, 10],
    [ 2,  8, 40],
]
labels = ["Class A", "Class B", "Class C"]
cm_df = pd.DataFrame(data, index=labels, columns=labels)

# 2) Plot raw counts
fig1, ax1 = plot(
    "confusion_matrix",
    cm=cm_df,
    labels=labels,  # optional, can be omitted since already in the DataFrame
    figsize=(6, 5),
    cmap="Oranges",
    normalize=False,
    fmt="d",
    cbar=True,
    save="confusion_matrix_raw"
)
plt.show()