Visualization Module

API reference for NumbaSOM visualization functions.

plot_u_matrix

numbasom.plot_u_matrix(u_mat, fig_size=(4, 4), colormap='viridis_r')

Plots the U-matrix in matplotlib.

Example: Plotting a U-Matrix

from numbasom import SOM, u_matrix, plot_u_matrix
import numpy as np

# Train a SOM
data = np.random.random([200, 3])
som = SOM(som_size=(50, 100))
lattice = som.train(data, num_iterations=15000)

# Compute and plot U-matrix
um = u_matrix(lattice)
plot_u_matrix(um, fig_size=(6.2, 6.2))

Parameters

• u_mat: The U-matrix computed from a trained lattice using u_matrix()
• fig_size: Tuple specifying the figure size (width, height) in inches. Default is (4, 4)
• colormap: Matplotlib colormap name. Default is 'viridis_r' (reversed viridis)

Available Colormaps

You can use any matplotlib colormap. Some good options for U-matrices:

• 'viridis_r' (default) - perceptually uniform, good for colorblind viewers
• 'plasma_r' - similar to viridis with more contrast
• 'coolwarm' - diverging colormap
• 'gray_r' - simple grayscale