Embedding Visualization

The fancy_embedding_pro() function creates publication-quality UMAP/t-SNE scatter plots with automatic label placement, density overlays, and equal-aspect axes.

Loading data

This tutorial uses the PBMC 3k dataset from Scanpy:

import scanpy as sc
adata = sc.datasets.pbmc3k_processed()

Basic categorical plot

Color cells by cluster identity with density contours:

from sjanpy.pl import fancy_embedding_pro

fancy_embedding_pro(adata, basis='umap', color='louvain')

This produces a scatter plot with:

Each cluster colored by the tab20 palette
KDE density contours in the background
Bold centroid labels with automatic repelling to avoid overlap
A legend on the right side

Customizing the plot

Change the legend title, hide density, and adjust dot size:

fancy_embedding_pro(
    adata,
    basis='umap',
    color='louvain',
    legend_title='Cell Type',
    show_density=False,
    dot_size=8,
    alpha=0.6,
    figsize=(12, 10),
)

Continuous variable (gene expression)

Pass a gene name to color to visualize expression:

fancy_embedding_pro(
    adata,
    basis='umap',
    color='CST3',
    legend_title='CST3 Expression',
    show_density=False,
)

When color is a gene name, the function automatically switches to a continuous colormap (viridis) and replaces the legend with a colorbar.

Saving the figure

fancy_embedding_pro(
    adata,
    basis='umap',
    color='louvain',
    save_path='embedding.pdf',
    dpi=300,
)