Skip to content



Bases: BaseArrowLayer

The PointCloudLayer renders a point cloud with 3D positions, normals and colors.

The PointCloudLayer can be more efficient at rendering large quantities of points than the ScatterplotLayer, but has fewer rendering options. In particular, you can have only one point size across all points in your data.


From GeoPandas:

import geopandas as gpd
from lonboard import Map, PointCloudLayer

# A GeoDataFrame with Point geometries
gdf = gpd.GeoDataFrame()
layer = PointCloudLayer.from_geopandas(
    get_color=[255, 0, 0],
m = Map(layer)

auto_highlight class-attribute instance-attribute

auto_highlight = tag(sync=True)

When true, the current object pointed to by the mouse pointer (when hovered over) is highlighted with highlightColor.

Requires pickable to be True.

  • Type: bool
  • Default: False

extensions class-attribute instance-attribute

extensions = tag(sync=True, **widget_serialization)

A list of layer extension objects to add additional features to a layer.

get_color class-attribute instance-attribute

get_color = ColorAccessor(None, allow_none=True)

The color of each path in the format of [r, g, b, [a]]. Each channel is a number between 0-255 and a is 255 if not supplied.

  • Type: ColorAccessor, optional
    • If a single list or tuple is provided, it is used as the color for all paths.
    • If a numpy or pyarrow array is provided, each value in the array will be used as the color for the path at the same row index.
  • Default: [0, 0, 0, 255].

get_normal class-attribute instance-attribute

get_normal = NormalAccessor(None, allow_none=True)

The normal of each object, in [nx, ny, nz].

  • Type: NormalAccessor, optional
    • If a single list or tuple is provided, it is used as the normal for all points.
    • If a numpy or pyarrow array is provided, each value in the array will be used as the normal for the point at the same row index.
  • Default: 1.

opacity class-attribute instance-attribute

opacity = tag(sync=True)

The opacity of the layer.

  • Type: float. Must range between 0 and 1.
  • Default: 1

pickable class-attribute instance-attribute

pickable = tag(sync=True)

Whether the layer responds to mouse pointer picking events.

This must be set to True for tooltips and other interactive elements to be available. This can also be used to only allow picking on specific layers within a map instance.

Note that picking has some performance overhead in rendering. To get the absolute best rendering performance with large data (at the cost of removing interactivity), set this to False.

  • Type: bool
  • Default: True

point_size class-attribute instance-attribute

point_size = tag(sync=True)

Global radius of all points, in units specified by size_units.

  • Type: float, optional
  • Default: 10

selected_index class-attribute instance-attribute

selected_index = tag(sync=True)

The positional index of the most-recently clicked on row of data.

You can use this to access the full row of data from a GeoDataFrame


Setting a value here from Python will do nothing. This attribute only exists to be updated from JavaScript on a map click. Note that pickable must be True (the default) on this layer for the JavaScript onClick handler to work; if pickable is set to False, selected_index will never update.

Note that you can use observe to call a function whenever a new value is received from JavaScript. Refer here for an example.

size_units class-attribute instance-attribute

size_units = tag(sync=True)

The units of the line width, one of 'meters', 'common', and 'pixels'. See unit system.

  • Type: str, optional
  • Default: 'pixels'

table class-attribute instance-attribute

table = PyarrowTableTrait(
    allowed_geometry_types={POINT}, allowed_dimensions={3}

A GeoArrow table with a Point column.

This is the fastest way to plot data from an existing GeoArrow source, such as geoarrow-rust or geoarrow-pyarrow.

If you have a GeoPandas GeoDataFrame, use from_geopandas instead.

visible class-attribute instance-attribute

visible = tag(sync=True)

Whether the layer is visible.

Under most circumstances, using the visible attribute to control the visibility of layers is recommended over removing/adding the layer from the Map.layers list.

In particular, toggling the visible attribute will persist the layer on the JavaScript side, while removing/adding the layer from the Map.layers list will re-download and re-render from scratch.

  • Type: bool
  • Default: True

from_duckdb classmethod

    sql: Union[str, DuckDBPyRelation],
    con: Optional[DuckDBPyConnection] = None,
    crs: Optional[Union[str, CRS]] = None,
    **kwargs: Unpack[PointCloudLayerKwargs]
) -> Self

Construct a Layer from a duckdb-spatial query.

DuckDB Spatial does not currently expose coordinate reference system information, so the user must ensure that data has been reprojected to EPSG:4326 or pass in the existing CRS of the data in the crs keyword parameter.


  • sql (Union[str, DuckDBPyRelation]) –

    The SQL input to visualize. This can either be a string containing a SQL query or the output of the duckdb sql function.

  • con (Optional[DuckDBPyConnection], default: None ) –

    The current DuckDB connection. This is required when passing a str to the sql parameter or when using a non-global DuckDB connection. Defaults to None.

Other Parameters:

  • crs (Optional[Union[str, CRS]]) –

    The CRS of the input data. This can either be a string passed to pyproj.CRS.from_user_input or a pyproj.CRS object. Defaults to None.


  • Self

    A Layer with the initialized data.

from_geopandas classmethod

    gdf: GeoDataFrame,
    auto_downcast: bool = True,
    **kwargs: Unpack[PointCloudLayerKwargs]
) -> Self

Construct a Layer from a geopandas GeoDataFrame.

The GeoDataFrame will be reprojected to EPSG:4326 if it is not already in that coordinate system.


  • gdf (GeoDataFrame) –

    The GeoDataFrame to set on the layer.

Other Parameters:


  • Self

    A Layer with the initialized data.