COG Layer Example¶

This notebook demonstrates how to visualize Cloud Optimized GeoTIFFs (COGs) using the @developmentseed/deck.gl-geotiff package.

The COG layer provides GPU-accelerated rendering with automatic reprojection support, allowing you to visualize large raster datasets directly in the browser without server-side processing.

MapLibre Example¶

The add_cog_layer method is available on the MapLibre Map class. It uses deck.gl overlay to render COG files.

In [ ]:

from anymap_ts import Map

# NLCD 2024 Land Cover COG (Continental US)
COG_URL = "https://s3.us-east-1.amazonaws.com/ds-deck.gl-raster-public/cog/Annual_NLCD_LndCov_2024_CU_C1V1.tif"

# Create a MapLibre map
m = Map(center=[-98.5, 39.8], zoom=4, style="dark-matter")

# Add the COG layer - the map will automatically fit to the COG bounds
m.add_cog_layer(COG_URL, name="nlcd-landcover", opacity=1.0)

m

from anymap_ts import Map # NLCD 2024 Land Cover COG (Continental US) COG_URL = "https://s3.us-east-1.amazonaws.com/ds-deck.gl-raster-public/cog/Annual_NLCD_LndCov_2024_CU_C1V1.tif" # Create a MapLibre map m = Map(center=[-98.5, 39.8], zoom=4, style="dark-matter") # Add the COG layer - the map will automatically fit to the COG bounds m.add_cog_layer(COG_URL, name="nlcd-landcover", opacity=1.0) m

DeckGL Example¶

The DeckGL map class also supports COG layers with additional debug options.

In [ ]:

from anymap_ts import DeckGLMap

# Create a DeckGL map
m2 = DeckGLMap(center=[-98.5, 39.8], zoom=4, style="dark-matter")

# Add the COG layer with debug mode to visualize the reprojection mesh
m2.add_cog_layer(
    COG_URL,
    name="nlcd-landcover-debug",
    debug=True,  # Show the reprojection mesh
    debug_opacity=0.25,
)

m2

from anymap_ts import DeckGLMap # Create a DeckGL map m2 = DeckGLMap(center=[-98.5, 39.8], zoom=4, style="dark-matter") # Add the COG layer with debug mode to visualize the reprojection mesh m2.add_cog_layer( COG_URL, name="nlcd-landcover-debug", debug=True, # Show the reprojection mesh debug_opacity=0.25, ) m2

Mapbox Example¶

The Mapbox map class also supports COG layers (requires a Mapbox access token).

In [ ]:

import os
from anymap_ts import MapboxMap

# Create a Mapbox map (requires MAPBOX_TOKEN environment variable)
# Skip if no token is set
if os.environ.get("MAPBOX_TOKEN"):
    m3 = MapboxMap(
        center=[-98.5, 39.8], zoom=4, style="mapbox://styles/mapbox/dark-v11"
    )

    # Add the COG layer
    m3.add_cog_layer(COG_URL, name="nlcd-landcover-mapbox", opacity=0.9)

    display(m3)
else:
    print("Set MAPBOX_TOKEN environment variable to use Mapbox maps")

import os from anymap_ts import MapboxMap # Create a Mapbox map (requires MAPBOX_TOKEN environment variable) # Skip if no token is set if os.environ.get("MAPBOX_TOKEN"): m3 = MapboxMap( center=[-98.5, 39.8], zoom=4, style="mapbox://styles/mapbox/dark-v11" ) # Add the COG layer m3.add_cog_layer(COG_URL, name="nlcd-landcover-mapbox", opacity=0.9) display(m3) else: print("Set MAPBOX_TOKEN environment variable to use Mapbox maps")

Configuration Options¶

The add_cog_layer method supports several configuration options:

Parameter	Type	Default	Description
url	str	required	URL to the Cloud Optimized GeoTIFF file
name	str	auto	Layer identifier
opacity	float	1.0	Layer opacity (0-1)
visible	bool	True	Whether layer is visible
debug	bool	False	Show reprojection mesh for debugging
debug_opacity	float	0.25	Opacity of debug mesh
max_error	float	0.125	Maximum reprojection error in pixels
fit_bounds	bool	True	Fit map to COG bounds after loading
before_id	str	None	ID of layer to insert before

About the Dataset¶

The example uses the NLCD 2024 Land Cover dataset:

• Coverage: Continental United States
• Resolution: 30 meters
• Source: USGS National Land Cover Database
• Format: Cloud Optimized GeoTIFF (COG)

The dataset classifies land cover into categories including:

• Developed areas (open space, low/medium/high intensity)
• Forest (deciduous, evergreen, mixed)
• Agricultural (pasture, crops)
• Wetlands
• Water bodies
• And more...

In [ ]:

# Export to standalone HTML
m.to_html("cog_layer_example.html")
print("Exported to cog_layer_example.html")

# Export to standalone HTML m.to_html("cog_layer_example.html") print("Exported to cog_layer_example.html")