Fast, interactive geospatial data visualization in Jupyter.

Building on cutting-edge technologies like GeoArrow and GeoParquet in conjunction with GPU-based map rendering, Lonboard enables visualizing large geospatial datasets interactively through a simple interface.

3 million points rendered from a GeoPandas GeoDataFrame in JupyterLab. Example notebook.

To install Lonboard using pip:

pip install lonboard

Lonboard is on conda-forge and can be installed using conda, mamba, or pixi. To install Lonboard using conda:

conda install -c conda-forge lonboard

To install from source, refer to the developer documentation.

For the simplest rendering, pass geospatial data into the top-level viz function.

import geopandas as gpd
from lonboard import viz

gdf = gpd.GeoDataFrame(...)
viz(gdf)

Under the hood, this delegates to a ScatterplotLayer, PathLayer, or PolygonLayer. Refer to the documentation and examples for more control over rendering.

Refer to the documentation at developmentseed.org/lonboard.

This is a new binding to the deck.gl geospatial data visualization library. A "deck" is the part of a skateboard you ride on. What's a fast, geospatial skateboard? A lonboard.

Illustration by Gus Becker.