Skip to content
Written for v1.5.0· Last updated: Oct 15, 2026

Facet View ​

Core Feature

This feature is available in all Portrai Explorer deployments.

Facet View splits the Map Viewer into a grid of small panels — one per category of a chosen categorical feature. This "small multiples" (or trellis) layout lets you compare how a single feature's pattern differs across cell types, clusters, samples, or any other categorical grouping, all at a glance.

It mirrors the split.by option in Seurat and the trellis layout in Tableau: the facet dimension (which feature splits the panels) is independent of the color dimension (the active color mapping). You can color by a continuous gene expression value and still split the view into per-category panels.

Enabling Facet View ​

Facet View is controlled from the Map Viewer toolbar.

  1. In the Explorer workspace, locate the Facet view button (grid icon) in the toolbar
  2. Click it to open the Facet by popover
  3. Under Facet by, select a categorical column to split by
  4. The workspace switches to the faceted grid

Selecting None returns the workspace to the standard single Map Viewer.

TIP

Only categorical features can be used as the facet dimension. Continuous features (gene expression, scores) are not listed in the Facet by selector — use them for color mapping instead.

WARNING

Facet View requires an active color mapping. If no feature or signature is set as the color basis, the grid shows the message "Color by a feature column to use the facet view." instead of panels — select a feature in the Explorer sidebar first.

Selecting Categories to Display ​

When faceting is active, a Categories to display checklist appears in the popover. Use it to restrict the grid to a subset of categories:

  • Only the checked categories are rendered as panels in the grid.
  • By default, all categories are shown.
  • If you uncheck every category, the grid shows a "Select at least one category to display" hint.

This is useful when a feature has many categories but you only want to compare a few.

Adjusting Panels per Row ​

When faceting is active, a Panels per row control in the popover sets how many panels appear in each row of the grid:

  • Auto (the default) fits the number of panels per row to the available width, reflowing as you resize the workspace.
  • Toggling Auto off — or dragging the Panels per row slider — fixes the grid to a chosen number of panels per row. The slider ranges from 1 to 12, and is capped at the number of currently visible panels (there is no point asking for more panels per row than there are panels).

The panels-per-row setting only affects the grid layout; it does not change which panels are shown.

How Panels Are Rendered ​

Each panel shows only the points belonging to that category that pass your active Subset, the cutoff, and any lasso or brush selection. A panel holds the points that match both of:

(points that pass the Subset, the cutoff and the lasso or brush selection)
AND (points in this category)

All panels share a single, consistent color scale derived from the active color mapping, so colors are directly comparable across panels — the recommended practice for small-multiple views. Each panel is labeled with its category name and the number of points it contains. All selected categories are shown at once in a single scrollable grid — scroll vertically to reach every panel. There is no limit on the total number of panels (the Panels per row control caps only how many appear per row, not how many are rendered overall).

Hovering a value in the sidebar (for example, a category in the Detail Panel) shows only the matching points in every panel simultaneously, until you move the pointer away.

Facet panels hide unselected points instead of dimming them. On the Map Viewer and the scatter plots, points a lasso or a brush left out stay on screen at reduced opacity; in a panel they are removed, so a panel always contains exactly the points its count reports.

Focusing on a Single Category ​

To explore one category on its own, hover over its panel and click the expand button (the maximize icon) that appears in the panel's top-right corner. A confirmation dialog asks whether to add that category as a subset condition and switch to the full-screen view.

On confirming:

  • The category is added to your active Subset as a condition on the facet feature, combined with AND so it narrows whatever is already filtered.
  • Facet View is turned off and the workspace returns to the standard single Map Viewer.
  • The map then shows only the points in that category.

The expand button appears only on panels that still contain points under your current filters — a panel that is already empty offers no drill-down, since focusing it would show nothing.

To return to the faceted grid, remove the added condition in the Subset Panel and re-enable Facet View.

TIP

Because the category becomes an ordinary subset condition, you can keep refining from there — add more conditions, draw a lasso, or apply a cutoff on top of the focused category.

Spatial and Embedding Projections ​

Facet View works with whichever projection is selected in the projection selector, which stays in the toolbar while faceting:

  • Spatial projections — the tissue image is captured as a background snapshot and aligned behind each panel, so points stay registered to the underlying image.
  • Embedding projections (UMAP, t-SNE) — panels are framed by the embedding's data extent, with no background image.

You can switch projections without leaving Facet View.

Exporting the Grid ​

The toolbar's Export button works in Facet View just as it does in the standard Map Viewer:

  • Capture PNG exports the entire facet grid as a single image.
  • CSV export of the underlying projection data is also available.

See Image Export for details.

  • Map Viewer - The standard single-view visualization
  • Features - Categorical features used as the facet dimension
  • Color Mapping - Configure the shared color scale
  • Subset - Control which points appear in each panel
  • Image Export - Export the facet grid as an image

Portrai Explorer Documentation