Styling Layers¶

The Style panel on the right edits the appearance of the layer selected in the Layers panel. Vector and raster layers each get their own set of controls.

Vector styling¶

For vector layers the Style panel covers fill, stroke, points, labels, and 3D extrusion:

• Fill: fill color and fill opacity for polygons.
• Stroke: line color and width for lines and polygon outlines.
• Points: circle radius for point layers.
• Labels: text color, size, halo color, and halo width.
• 3D extrusion: turn polygons into extruded blocks, with a height field, height scale, base height, and color. Advanced expressions are available for both height and color.

You can also set a per-style minimum and maximum zoom so a style only applies within a zoom range.

Point renderer (heatmap and clustering)¶

For point-only GeoJSON layers — whether dropped on the map, produced by a tool, or loaded through Add Vector Layer in the geojson render mode — the Style panel adds a Point renderer control:

The renderer choice is saved with the project.

Style type (data-driven styling)¶

The Style type control chooses how feature values map to color:

Graduated and categorized styles share four more controls, and then list the generated stops so you can fine-tune an individual colour before applying:

• Attribute — the field the style reads.
• Classes — 2 to 12 classes for a graduated style; 1 to 12 for a categorized one, plus an All (n) option that gives every distinct value its own colour.
• Scheme — how the stops are chosen. Graduated offers Equal interval, Quantile, and Natural breaks; categorized offers Most frequent, Alphabetical, and First values.
• Colormap — the named colour ramp the classes are drawn from.

Nothing reaches the map until you click Apply style type, so you can adjust the classification and watch the stop list update first.

Diagram symbology¶

Below the renderer controls, a Diagram control draws a small chart on each feature instead of (or as well as) a plain symbol: Pie chart, Donut chart, Bar chart, or Stacked bar chart, built from the numeric fields you pick. Set it to None to turn diagrams off. It is the usual way to show a composition — vote share, land-cover mix, age structure — per polygon or per point.

Quick filters¶

Below the renderer and label controls, the Quick filters section narrows what the layer draws without writing an expression. Pick a field from Add a filter... and GeoLibre reads the data to decide which control it deserves:

• a Values list of checkboxes with a count beside each value (with a search box once there are more than a handful),
• a Range slider plus typed minimum and maximum for a numeric field,
• a Dates from/to pair for a timestamp field, with both days included in full,
• a Text match — contains, starts with, or is exactly — which ignores case.

The Filter type dropdown beside a field switches between the controls it supports, so a numeric code column that opens as a range can be answered with checkboxes instead. Several filters on one layer narrow it together, and they combine with a Time Slider window, an embed setFilter, and rule-based symbology rather than replacing them.

A quick filter hides features; it does not select them. Use Select by Expression or the selection tools when you want to act on features rather than take them off the map.

A layer with an active filter shows a funnel icon on its row in the Layers panel, so a filtered layer is never mistaken for missing data, and its Layer actions -> Clear filters empties every control at once while keeping the controls themselves. Clearing the last checkbox in a Values list is the same as clearing that filter: it shows every value, not none. Remove all at the top of the section deletes the controls outright.

Filters are saved with the project, so a shared map opens with the same view of the data. They also appear in the read-only viewer chrome (?layout=viewer) — filtering is a way of reading a map, so the person you shared it with can ask it questions even though the authoring panels are hidden.

Tile-backed layers (vector tiles, PMTiles, MBTiles) carry no local features, so their value lists are read from the tiles currently loaded and grow as you pan and zoom; the section says so under the controls.

One limit is worth knowing: a point layer using the cluster renderer builds its clusters from the layer's whole dataset before any filter applies, so while clustering is on the bubbles and their counts describe the unfiltered data. The same is true of a Time Slider window and a rule-based filter. Switch the point renderer back to Single symbol to see the filtered features drawn individually, one circle each, with no cluster counts to misread.

Style interchange and URL styles¶

The selected vector layer's Layer actions → Styles submenu imports and exports GeoLibre URL style JSON, Mapbox/MapLibre style JSON, OGC SLD, and QGIS QML. A GeoLibre URL style is a compact MapLibre style whose feature data is supplied separately through the data URL parameter. Export one when you want to publish the current symbology beside hosted GeoJSON or a ZIP of GeoJSON files; import it when you want to apply that symbology to a layer already open in GeoLibre.

See Managing Layers for the menu workflow and Embedding & Sharing for the JSON conventions.

Popups and hover tooltips¶

Further down the Style panel, the Popup section designs what a viewer sees when they interact with a feature, rather than the raw column dump the Identify popup shows by default. It matters most for a map you share: a layout=viewer embed, a shared project, or a story-map chapter, where the recipient clicks a feature and would otherwise get join artifacts, editor-tracking columns, and internal ids alongside the fields you meant them to read.

• Show popup on click (Identify) and Show tooltip on hover are independent switches. The click popup is on by default; the hover tooltip is off.
• Title field leads the popup with the feature's own name rather than the layer's. Title expression does the same from an expression, and wins over the field.
• Body expression replaces the whole popup body — the field rows and the feature id row with them — with a sentence built from the feature's properties, for authors who want prose rather than a table. Both boxes open the same expression builder that backs data-driven styling, labels, and filters.
• Fields chooses which attributes appear and in what order. Add a field from the picker, drag its handle (or use the arrow buttons) to reorder, and give it a Label to show instead of the raw column name. Leave the list empty — or remove every field you added — to keep the default: every visible field, in the data's own order.
• Value type formats the value: Number (decimals, thousands separator), Date (date, date and time, time, ISO 8601, or year), Link, Image, or plain Text. For Text, Number and Date fields, Prefix and Suffix wrap the result, which is where a currency symbol or a unit goes.
• In hover tooltip puts a field in the tooltip's short subset. Keep it to one or two fields: the tip follows the pointer, so it is a glance, not a table. Image fields are skipped there and stay in the click popup, where there is room to show the picture.

Layers loaded through Add Vector Layer are designed here too. That panel has its own Popup checkbox per layer, which opens a separate, unstyled attribute table on click; it is off by default so the two do not both answer the same click, and the design above is what a viewer gets. Tick it only if you want the control's raw table back.

Fields you hid or excluded from the attribute table stay out of the popup: the popup design selects from the visible fields and cannot re-expose a hidden one. A layer with nothing configured behaves exactly as before, and the whole design saves with the project, so it travels to the Python package and the MCP authoring tools as well.

Raster styling¶

For raster layers the Style panel exposes image adjustments:

• Brightness (minimum and maximum)
• Saturation
• Contrast
• Hue rotation (in degrees)

These let you tune the look of GeoTIFF, COG, and tile-based raster layers without changing the underlying data.

Spectral profile¶

For a multiband raster — a stacked Landsat or Sentinel scene, a NetCDF/HDF cube, or any COG with more than one band — the Style panel adds a Spectral profile chart of one pixel's value across every band.

Turn on Identify for the layer in the Layers panel, then click the map. Each click adds a numbered dot on the map and a matching curve in the chart, so you can click water, vegetation, and asphalt and compare the three responses side by side. Up to six points are compared at once; older points age off as you add more.

• The chart plots against wavelength when the file declares one wavelength per band, and against band number otherwise. A wavelength list that doesn't match the band count is ignored rather than trusted, so a stale one can't mislabel the axis.
• Only the tile containing the clicked pixel is fetched, not the whole scene, so profiling a large remote COG stays fast.
• Pop out floats the chart in a draggable, resizable window over the map; PNG and CSV export it.

Single-band rasters produce no profile — use Identify on its own to read the pixel value. Clicks outside the raster, or on a pixel that is nodata in every band, are dropped without disturbing the points you have already collected.

A remote GeoJSON, GeoParquet, or vector PMTiles layer opened with the data URL parameter can receive a GeoLibre/MapLibre vector style through style. A remote COG can receive a raster style instead. Raster URL styles support band mode and selection, rescale ranges, colormap and reversal, nodata, opacity, gamma, stretch, and normalized-difference index presets. See the remote data examples.

Legends and colorbars¶

To display a legend or a continuous colorbar on the map, open them from the Controls menu. They reflect the styling you set here.