Setup

Overview

Use the notebook below on a Databricks cluster to download the Essential and Complete sample-data bundles into a Unity Catalog Volume. The notebook uses the GeoBrix sample module (packaged with the GeoBrix WHL), so you only need to install GeoBrix and run the cells.

Prerequisites

Unity Catalog Volume must already exist: {CATALOG}.{SCHEMA}.{VOLUME}
Create volume in Databricks UI or via SQL: CREATE VOLUME IF NOT EXISTS {CATALOG}.{SCHEMA}.{VOLUME}
Volumes paths (/Volumes/...) work directly with GDAL, rasterio, and other libraries
No /dbfs translation needed in modern Databricks

Databricks vs local use

This configuration is intended for use with Databricks Unity Catalog volumes. The path format /Volumes/<catalog>/<schema>/<volume>/... can also be mimicked for local development—for example, use a local path such as ./Volumes/main/default/geobrix_samples/geobrix-examples and set the base path to that so the same code and examples work outside Databricks.

Notebook

Run the cells in order. The notebook uses databricks.labs.gbx.sample (included in the GeoBrix package): get_volumes_path, run_essential_bundle, and run_complete_bundle. Skips datasets that already exist; uses a temp directory for interim downloads; copies only final files to the Volume.

Download notebook

setup_sample_data.ipynb — download and import into your Databricks workspace to run the setup.

GeoBrix Sample Data Setup

Run this notebook on a Databricks cluster to download the Essential and Complete sample-data bundles into a Unity Catalog Volume. Uses the GeoBrix `sample` module (included in the GeoBrix WHL)—install GeoBrix and run the cells in order.

- Skips datasets that already exist at the Volumes path
Uses a local temp directory for interim downloads (zip, convert); only final files are copied to the Volume

1. Install dependencies (once per session)

Install dependencies

%pip install -q requests pystac-client planetary-computer geopandas

2. Config: catalog, schema, volume

from databricks.labs.gbx.sample import get_volumes_path, run_essential_bundle, run_complete_bundle
from pathlib import Path

CATALOG = "main"
SCHEMA = "default"
VOLUME = "geobrix_samples"
SAMPLE_DATA_PATH = get_volumes_path(CATALOG, SCHEMA, VOLUME)
print(f"SAMPLE_DATA_PATH = {SAMPLE_DATA_PATH}")

## -- uncomment to remove everything under 'geobrix-examples' subfolder
#dbutils.fs.rm(SAMPLE_DATA_PATH, recurse=True)

3. Run Essential Bundle (~355 MB)

NYC & London vectors (GeoJSON), Sentinel-2 rasters, SRTM elevation. Skips files that already exist.

try:
    Path(SAMPLE_DATA_PATH).mkdir(parents=True, exist_ok=True)
except Exception:
    print("... VOLUME path must already exist.")
    pass
result_essential = run_essential_bundle(SAMPLE_DATA_PATH)
print(f"Files: {result_essential['file_count']}, Total: {result_essential['total_size_mb']:.1f} MB")
if result_essential["errors"]:
    for name, err in result_essential["errors"]:
        print(f"  ⚠️ {name}: {err[:80]}...")

4. Run Complete Bundle (~185 MB more)

Neighborhoods, extra elevation, HRRR weather, shapefiles (parks, subway), multi-layer GeoPackage. Run after Essential.

result_complete = run_complete_bundle(SAMPLE_DATA_PATH)
print(f"Files: {result_complete['file_count']}, Total: {result_complete['total_size_mb']:.1f} MB")
if result_complete["errors"]:
    for name, err in result_complete["errors"]:
        print(f"  ⚠️ {name}: {err[:80]}...")

5. Summary

print(f"Sample data base path: {SAMPLE_DATA_PATH}")
for item in sorted(Path(SAMPLE_DATA_PATH).rglob("*")):
    if item.is_file():
        rel = item.relative_to(SAMPLE_DATA_PATH)
        print(f"  {rel}: {item.stat().st_size / (1024*1024):.1f} MB")

Output

After running the notebook, the Volume (or your configured base path) contains the following files under geobrix-examples/. The table is ordered by bundle (Essential, then Complete) and within each by region (NYC, then London).

Bundle	Source	Path	Size	Format & use
Essential	NYC Boroughs	`nyc/boroughs/nyc_boroughs.geojson`	3.0 MB	GeoJSON — vector boundaries; joins, overlay, mapping. Driver: GeoJSON
Essential	NYC Taxi Zones	`nyc/taxi-zones/nyc_taxi_zones.geojson`	3.7 MB	GeoJSON — vector zones; spatial joins, aggregation. Driver: GeoJSON
Essential	SRTM NYC	`nyc/elevation/srtm_n40w074.tif`	24.7 MB	GeoTIFF — elevation DEM; slope, aspect, hillshade. Driver: GTiff
Essential	NYC Sentinel-2	`nyc/sentinel2/nyc_sentinel2_red.tif`	205.4 MB	GeoTIFF — optical raster (red band); band math, indices. Driver: GTiff
Essential	London Postcodes	`london/postcodes/london_postcodes.geojson`	0.9 MB	GeoJSON — vector postcode areas; point-in-polygon, mapping. Driver: GeoJSON
Essential	SRTM London	`london/elevation/srtm_n51w001.tif`	24.7 MB	GeoTIFF — elevation DEM; terrain analysis. Driver: GTiff
Essential	London Sentinel-2	`london/sentinel2/london_sentinel2_red.tif`	92.7 MB	GeoTIFF — optical raster; NDVI, classification. Driver: GTiff
Complete	NYC Neighborhoods	`nyc/neighborhoods/nyc_nta.geojson`	4.1 MB	GeoJSON — vector neighborhoods (NTA); aggregation, mapping. Driver: GeoJSON
Complete	SRTM NYC East	`nyc/elevation/srtm_n40w073.tif`	24.7 MB	GeoTIFF — elevation DEM; extended terrain. Driver: GTiff
Complete	HRRR Weather	`nyc/hrrr-weather/hrrr_nyc_20260213_12z.grib2`	131.2 MB	GRIB2 — weather model; temperature, wind, precipitation. Driver: GRIB
Complete	NYC Parks	`nyc/parks/nyc_parks.shp.zip`	2.1 MB	Shapefile (zip) — vector polygons; overlay, area stats. Driver: ESRI Shapefile (use `/vsizip/` path)
Complete	NYC Subway	`nyc/subway/nyc_subway.shp.zip`	0.1 MB	Shapefile (zip) — vector points (stations); proximity, routing. Driver: ESRI Shapefile (use `/vsizip/` path)
Complete	Multi-Layer GeoPackage	`nyc/geopackage/nyc_complete.gpkg`	7.0 MB	GeoPackage — multi-layer vector (boroughs, zones, parks, subway); single-file DB. Driver: GPKG
Complete	London Boroughs	`london/boroughs/london_boroughs.geojson`	1.9 MB	GeoJSON — vector boundaries; joins, overlay. Driver: GeoJSON

Exact sizes may vary slightly; the HRRR filename date reflects the run date at download time.

Overview​

Notebook​