GeoMaster
Comprehensive geospatial science skill covering GIS, remote sensing, spatial analysis, and ML for Earth observation across 70+ topics with 500+ code examples in 8 programming languages.
Installation
# Core Python stack (conda recommended)
conda install -c conda-forge gdal rasterio fiona shapely pyproj geopandas
# Remote sensing & ML
conda install -c conda-forge rsgislib # rsgislib is conda-only, not on PyPI
uv pip install torchgeo earthengine-api
uv pip install scikit-learn xgboost torch-geometric
# Network & visualization
uv pip install osmnx networkx folium keplergl
uv pip install cartopy contextily mapclassify
# Big data & cloud
uv pip install xarray rioxarray dask-geopandas
uv pip install pystac-client planetary-computer
# Point clouds (pylas is deprecated upstream in favor of laspy 2.0+; use laspy only)
uv pip install laspy open3d pdal
# Databases
conda install -c conda-forge postgis spatialite
Quick Start
NDVI from Sentinel-2
import rasterio
import numpy as np
with rasterio.open('sentinel2.tif') as src:
red = src.read(4).astype(float) # B04
nir = src.read(8).astype(float) # B08
ndvi = (nir - red) / (nir + red + 1e-8)
ndvi = np.nan_to_num(ndvi, nan=0)
profile = src.profile
profile.update(count=1, dtype=rasterio.float32)
with rasterio.open('ndvi.tif', 'w', **profile) as dst:
dst.write(ndvi.astype(rasterio.float32), 1)
Spatial Analysis with GeoPandas
import geopandas as gpd
# Load and ensure same CRS
zones = gpd.read_file('zones.geojson')
points = gpd.read_file('points.geojson')
if zones.crs != points.crs:
points = points.to_crs(zones.crs)
# Spatial join and statistics
joined = gpd.sjoin(points, zones, how='inner', predicate='within')
stats = joined.groupby('zone_id').agg({
'value': ['count', 'mean', 'std', 'min', 'max']
}).round(2)
Google Earth Engine Time Series
import ee
import pandas as pd
ee.Initialize(project='your-project')
roi = ee.Geometry.Point([-122.4, 37.7]).buffer(10000)
s2 = (ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED')
.filterBounds(roi)
.filterDate('2020-01-01', '2023-12-31')
.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 20)))
def add_ndvi(img):
return img.addBands(img.normalizedDifference(['B8', 'B4']).rename('NDVI'))
s2_ndvi = s2.map(add_ndvi)
def extract_series(image):
stats = image.reduceRegion(ee.Reducer.mean(), roi.centroid(), scale=10, maxPixels=1e9)
return ee.Feature(None, {'date': image.date().format('YYYY-MM-dd'), 'ndvi': stats.get('NDVI')})
series = s2_ndvi.map(extract_series).getInfo()
df = pd.DataFrame([f['properties'] for f in series['features']])
df['date'] = pd.to_datetime(df['date'])
Core Concepts
Data Types
| Type |
Examples |
Libraries |
| Vector |
Shapefile, GeoJSON, GeoPackage |
GeoPandas, Fiona, GDAL |
| Raster |
GeoTIFF, NetCDF, COG |
Rasterio, Xarray, GDAL |
| Point Cloud |
LAS, LAZ |
Laspy, PDAL, Open3D |
Coordinate Systems
- EPSG:4326 (WGS 84) - Geographic, lat/lon, use for storage
- EPSG:3857 (Web Mercator) - Web maps only (don't use for area/distance!)
- EPSG:326xx/327xx (UTM) - Metric calculations, <1% distortion per zone
- Use
gdf.estimate_utm_crs() for automatic UTM detection
# Always check CRS before operations
assert gdf1.crs == gdf2.crs, "CRS mismatch!"
# For area/distance calculations, use projected CRS
gdf_metric = gdf.to_crs(gdf.estimate_utm_crs())
area_sqm = gdf_metric.geometry.area
OGC Standards
- WMS: Web Map Service - raster maps
- WFS: Web Feature Service - vector data
- WCS: Web Coverage Service - raster coverage
- STAC: Spatiotemporal Asset Catalog - modern metadata
Common Operations
Spectral Indices
def calculate_indices(image_path):
"""NDVI, EVI, SAVI, NDWI from Sentinel-2."""
with rasterio.open(image_path) as src:
B02, B03, B04, B08, B11 = [src.read(i).astype(float) for i in [1,2,3,4,5]]
ndvi = (B08 - B04) / (B08 + B04 + 1e-8)
evi = 2.5 * (B08 - B04) / (B08 + 6*B04 - 7.5*B02 + 1)
savi = ((B08 - B04) / (B08 + B04 + 0.5)) * 1.5
ndwi = (B03 - B08) / (B03 + B08 + 1e-8)
return {'NDVI': ndvi, 'EVI': evi, 'SAVI': savi, 'NDWI': ndwi}
Vector Operations
# Buffer (use projected CRS!)
gdf_proj = gdf.to_crs(gdf.estimate_utm_crs())
gdf['buffer_1km'] = gdf_proj.geometry.buffer(1000)
# Spatial relationships
intersects = gdf[gdf.geometry.intersects(other_geometry)]
contains = gdf[gdf.geometry.contains(point_geometry)]
# Geometric operations
gdf['centroid'] = gdf.geometry.centroid
gdf['simplified'] = gdf.geometry.simplify(tolerance=0.001)
# Overlay operations
intersection = gpd.overlay(gdf1, gdf2, how='intersection')
union = gpd.overlay(gdf1, gdf2, how='union')
Terrain Analysis
def terrain_metrics(dem_path):
"""Calculate slope, aspect, hillshade from DEM."""
with rasterio.open(dem_path) as src:
dem = src.read(1)
dy, dx = np.gradient(dem)
slope = np.arctan(np.sqrt(dx**2 + dy**2)) * 180 / np.pi
aspect = (90 - np.arctan2(-dy, dx) * 180 / np.pi) % 360
# Hillshade
az_rad, alt_rad = np.radians(315), np.radians(45)
hillshade = (np.sin(alt_rad) * np.sin(np.radians(slope)) +
np.cos(alt_rad) * np.cos(np.radians(slope)) *
np.cos(np.radians(aspect) - az_rad))
return slope, aspect, hillshade
Network Analysis
import osmnx as ox
import networkx as nx
# Download and analyze street network
G = ox.graph_from_place('San Francisco, CA', network_type='drive')
G = ox.add_edge_speeds(G).add_edge_travel_times(G)
# Shortest path
orig = ox.distance.nearest_nodes(G, -122.4, 37.7)
dest = ox.distance.nearest_nodes(G, -122.3, 37.8)
route = nx.shortest_path(G, orig, dest, weight='travel_time')
Image Classification
from sklearn.ensemble import RandomForestClassifier
import rasterio
from rasterio.features import rasterize
def classify_imagery(raster_path, training_gdf, output_path):
"""Train RF and classify imagery."""
with rasterio.open(raster_path) as src:
image = src.read()
profile = src.profile
transform = src.transform
# Extract training data
X_train, y_train = [], []
for _, row in training_gdf.iterrows():
mask = rasterize([(row.geometry, 1)],
out_shape=(profile['height'], profile['width']),
transform=transform, fill=0, dtype=np.uint8)
pixels = image[:, mask > 0].T
X_train.extend(pixels)
y_train.extend([row['class_id']] * len(pixels))
# Train and predict
rf = RandomForestClassifier(n_estimators=100, max_depth=20, n_jobs=-1)
rf.fit(X_train, y_train)
prediction = rf.predict(image.reshape(image.shape[0], -1).T)
prediction = prediction.reshape(profile['height'], profile['width'])
profile.update(dtype=rasterio.uint8, count=1)
with rasterio.open(output_path, 'w', **profile) as dst:
dst.write(prediction.astype(rasterio.uint8), 1)
return rf
Modern Cloud-Native Workflows
STAC + Planetary Computer
import pystac_client
import planetary_computer
import odc.stac
# Search Sentinel-2 via STAC
catalog = pystac_client.Client.open(
"https://planetarycomputer.microsoft.com/api/stac/v1",
modifier=planetary_computer.sign_inplace,
)
search = catalog.search(
collections=["sentinel-2-l2a"],
bbox=[-122.5, 37.7, -122.3, 37.9],
datetime="2023-01-01/2023-12-31",
query={"eo:cloud_cover": {"lt": 20}},
)
# Load as xarray (cloud-native!)
data = odc.stac.load(
list(search.get_items())[:5],
bands=["B02", "B03", "B04", "B08"],
crs="EPSG:32610",
resolution=10,
)
# Calculate NDVI on xarray
ndvi = (data.B08 - data.B04) / (data.B08 + data.B04)
Cloud-Optimized GeoTIFF (COG)
import rasterio
from rasterio.session import AWSSession
# Read COG directly from cloud (partial reads)
session = AWSSession(aws_access_key_id=..., aws_secret_access_key=...)
with rasterio.open('s3://bucket/path.tif', session=session) as src:
# Read only window of interest
window = ((1000, 2000), (1000, 2000))
subset = src.read(1, window=window)
# Write COG
with rasterio.open('output.tif', 'w', **profile,
tiled=True, blockxsize=256, blockysize=256,
compress='DEFLATE', predictor=2) as dst:
dst.write(data)
# Validate COG
from rio_cogeo.cogeo import cog_validate
cog_validate('output.tif')
Performance Tips
# 1. Spatial indexing (10-100x faster queries)
gdf.sindex # Auto-created by GeoPandas
# 2. Chunk large rasters
with rasterio.open('large.tif') as src:
for i, window in src.block_windows(1):
block = src.read(1, window=window)
# 3. Dask for big data
import dask.array as da
dask_array = da.from_rasterio('large.tif', chunks=(1, 1024, 1024))
# 4. Use Arrow for I/O
gdf.to_file('output.gpkg', use_arrow=True)
# 5. GDAL caching
from osgeo import gdal
gdal.SetCacheMax(2**30) # 1GB cache
# 6. Parallel processing
rf = RandomForestClassifier(n_jobs=-1) # All cores
Best Practices
- Always check CRS before spatial operations
- Use projected CRS for area/distance calculations
- Validate geometries:
gdf = gdf[gdf.is_valid]
- Handle missing data:
gdf['geometry'] = gdf['geometry'].fillna(None)
- Use efficient formats: GeoPackage > Shapefile, Parquet for large data
- Apply cloud masking to optical imagery
- Preserve lineage for reproducible research
- Use appropriate resolution for your analysis scale
Detailed Documentation
- Coordinate Systems - CRS fundamentals, UTM, transformations
- Core Libraries - GDAL, Rasterio, GeoPandas, Shapely
- Remote Sensing - Satellite missions, spectral indices, SAR
- Machine Learning - Deep learning, CNNs, GNNs for RS
- GIS Software - QGIS, ArcGIS, GRASS integration
- Scientific Domains - Marine, hydrology, agriculture, forestry
- Advanced GIS - 3D GIS, spatiotemporal, topology
- Big Data - Distributed processing, GPU acceleration
- Industry Applications - Urban planning, disaster management
- Programming Languages - Python, R, Julia, JS, C++, Java, Go, Rust
- Data Sources - Satellite catalogs, APIs
- Troubleshooting - Common issues, debugging, error reference
- Code Examples - 500+ examples
GeoMaster covers everything from basic GIS operations to advanced remote sensing and machine learning.
1---2name: geomaster3description: Comprehensive geospatial science skill covering remote sensing, GIS, spatial analysis, machine learning for earth observation, and 30+ scientific domains. Supports satellite imagery processing (Sentinel, Landsat, MODIS, SAR, hyperspectral), vector and raster data operations, spatial statistics, point cloud processing, network analysis, cloud-native workflows (STAC, COG, Planetary Computer), and 8 programming languages (Python, R, Julia, JavaScript, C++, Java, Go, Rust) with 500+ code examples. Use for remote sensing workflows, GIS analysis, spatial ML, Earth observation data processing, terrain analysis, hydrological modeling, marine spatial analysis, atmospheric science, and any geospatial computation task.4license: MIT License5metadata:6 source: https://skillmd.com/skills/stanfish06/geomaster7 version: "1.0"8 skill-author: K-Dense Inc.9---1011# GeoMaster1213Comprehensive geospatial science skill covering GIS, remote sensing, spatial analysis, and ML for Earth observation across 70+ topics with 500+ code examples in 8 programming languages.1415## Installation1617```bash18# Core Python stack (conda recommended)19conda install -c conda-forge gdal rasterio fiona shapely pyproj geopandas2021# Remote sensing & ML22conda install -c conda-forge rsgislib # rsgislib is conda-only, not on PyPI23uv pip install torchgeo earthengine-api24uv pip install scikit-learn xgboost torch-geometric2526# Network & visualization27uv pip install osmnx networkx folium keplergl28uv pip install cartopy contextily mapclassify2930# Big data & cloud31uv pip install xarray rioxarray dask-geopandas32uv pip install pystac-client planetary-computer3334# Point clouds (pylas is deprecated upstream in favor of laspy 2.0+; use laspy only)35uv pip install laspy open3d pdal3637# Databases38conda install -c conda-forge postgis spatialite39```4041## Quick Start4243### NDVI from Sentinel-24445```python46import rasterio47import numpy as np4849with rasterio.open('sentinel2.tif') as src:50 red = src.read(4).astype(float) # B0451 nir = src.read(8).astype(float) # B0852 ndvi = (nir - red) / (nir + red + 1e-8)53 ndvi = np.nan_to_num(ndvi, nan=0)5455 profile = src.profile56 profile.update(count=1, dtype=rasterio.float32)5758 with rasterio.open('ndvi.tif', 'w', **profile) as dst:59 dst.write(ndvi.astype(rasterio.float32), 1)60```6162### Spatial Analysis with GeoPandas6364```python65import geopandas as gpd6667# Load and ensure same CRS68zones = gpd.read_file('zones.geojson')69points = gpd.read_file('points.geojson')7071if zones.crs != points.crs:72 points = points.to_crs(zones.crs)7374# Spatial join and statistics75joined = gpd.sjoin(points, zones, how='inner', predicate='within')76stats = joined.groupby('zone_id').agg({77 'value': ['count', 'mean', 'std', 'min', 'max']78}).round(2)79```8081### Google Earth Engine Time Series8283```python84import ee85import pandas as pd8687ee.Initialize(project='your-project')88roi = ee.Geometry.Point([-122.4, 37.7]).buffer(10000)8990s2 = (ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED')91 .filterBounds(roi)92 .filterDate('2020-01-01', '2023-12-31')93 .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 20)))9495def add_ndvi(img):96 return img.addBands(img.normalizedDifference(['B8', 'B4']).rename('NDVI'))9798s2_ndvi = s2.map(add_ndvi)99100def extract_series(image):101 stats = image.reduceRegion(ee.Reducer.mean(), roi.centroid(), scale=10, maxPixels=1e9)102 return ee.Feature(None, {'date': image.date().format('YYYY-MM-dd'), 'ndvi': stats.get('NDVI')})103104series = s2_ndvi.map(extract_series).getInfo()105df = pd.DataFrame([f['properties'] for f in series['features']])106df['date'] = pd.to_datetime(df['date'])107```108109## Core Concepts110111### Data Types112113| Type | Examples | Libraries |114|------|----------|-----------|115| Vector | Shapefile, GeoJSON, GeoPackage | GeoPandas, Fiona, GDAL |116| Raster | GeoTIFF, NetCDF, COG | Rasterio, Xarray, GDAL |117| Point Cloud | LAS, LAZ | Laspy, PDAL, Open3D |118119### Coordinate Systems120121- **EPSG:4326** (WGS 84) - Geographic, lat/lon, use for storage122- **EPSG:3857** (Web Mercator) - Web maps only (don't use for area/distance!)123- **EPSG:326xx/327xx** (UTM) - Metric calculations, <1% distortion per zone124- Use `gdf.estimate_utm_crs()` for automatic UTM detection125126```python127# Always check CRS before operations128assert gdf1.crs == gdf2.crs, "CRS mismatch!"129130# For area/distance calculations, use projected CRS131gdf_metric = gdf.to_crs(gdf.estimate_utm_crs())132area_sqm = gdf_metric.geometry.area133```134135### OGC Standards136137- **WMS**: Web Map Service - raster maps138- **WFS**: Web Feature Service - vector data139- **WCS**: Web Coverage Service - raster coverage140- **STAC**: Spatiotemporal Asset Catalog - modern metadata141142## Common Operations143144### Spectral Indices145146```python147def calculate_indices(image_path):148 """NDVI, EVI, SAVI, NDWI from Sentinel-2."""149 with rasterio.open(image_path) as src:150 B02, B03, B04, B08, B11 = [src.read(i).astype(float) for i in [1,2,3,4,5]]151152 ndvi = (B08 - B04) / (B08 + B04 + 1e-8)153 evi = 2.5 * (B08 - B04) / (B08 + 6*B04 - 7.5*B02 + 1)154 savi = ((B08 - B04) / (B08 + B04 + 0.5)) * 1.5155 ndwi = (B03 - B08) / (B03 + B08 + 1e-8)156157 return {'NDVI': ndvi, 'EVI': evi, 'SAVI': savi, 'NDWI': ndwi}158```159160### Vector Operations161162```python163# Buffer (use projected CRS!)164gdf_proj = gdf.to_crs(gdf.estimate_utm_crs())165gdf['buffer_1km'] = gdf_proj.geometry.buffer(1000)166167# Spatial relationships168intersects = gdf[gdf.geometry.intersects(other_geometry)]169contains = gdf[gdf.geometry.contains(point_geometry)]170171# Geometric operations172gdf['centroid'] = gdf.geometry.centroid173gdf['simplified'] = gdf.geometry.simplify(tolerance=0.001)174175# Overlay operations176intersection = gpd.overlay(gdf1, gdf2, how='intersection')177union = gpd.overlay(gdf1, gdf2, how='union')178```179180### Terrain Analysis181182```python183def terrain_metrics(dem_path):184 """Calculate slope, aspect, hillshade from DEM."""185 with rasterio.open(dem_path) as src:186 dem = src.read(1)187188 dy, dx = np.gradient(dem)189 slope = np.arctan(np.sqrt(dx**2 + dy**2)) * 180 / np.pi190 aspect = (90 - np.arctan2(-dy, dx) * 180 / np.pi) % 360191192 # Hillshade193 az_rad, alt_rad = np.radians(315), np.radians(45)194 hillshade = (np.sin(alt_rad) * np.sin(np.radians(slope)) +195 np.cos(alt_rad) * np.cos(np.radians(slope)) *196 np.cos(np.radians(aspect) - az_rad))197198 return slope, aspect, hillshade199```200201### Network Analysis202203```python204import osmnx as ox205import networkx as nx206207# Download and analyze street network208G = ox.graph_from_place('San Francisco, CA', network_type='drive')209G = ox.add_edge_speeds(G).add_edge_travel_times(G)210211# Shortest path212orig = ox.distance.nearest_nodes(G, -122.4, 37.7)213dest = ox.distance.nearest_nodes(G, -122.3, 37.8)214route = nx.shortest_path(G, orig, dest, weight='travel_time')215```216217## Image Classification218219```python220from sklearn.ensemble import RandomForestClassifier221import rasterio222from rasterio.features import rasterize223224def classify_imagery(raster_path, training_gdf, output_path):225 """Train RF and classify imagery."""226 with rasterio.open(raster_path) as src:227 image = src.read()228 profile = src.profile229 transform = src.transform230231 # Extract training data232 X_train, y_train = [], []233 for _, row in training_gdf.iterrows():234 mask = rasterize([(row.geometry, 1)],235 out_shape=(profile['height'], profile['width']),236 transform=transform, fill=0, dtype=np.uint8)237 pixels = image[:, mask > 0].T238 X_train.extend(pixels)239 y_train.extend([row['class_id']] * len(pixels))240241 # Train and predict242 rf = RandomForestClassifier(n_estimators=100, max_depth=20, n_jobs=-1)243 rf.fit(X_train, y_train)244245 prediction = rf.predict(image.reshape(image.shape[0], -1).T)246 prediction = prediction.reshape(profile['height'], profile['width'])247248 profile.update(dtype=rasterio.uint8, count=1)249 with rasterio.open(output_path, 'w', **profile) as dst:250 dst.write(prediction.astype(rasterio.uint8), 1)251252 return rf253```254255## Modern Cloud-Native Workflows256257### STAC + Planetary Computer258259```python260import pystac_client261import planetary_computer262import odc.stac263264# Search Sentinel-2 via STAC265catalog = pystac_client.Client.open(266 "https://planetarycomputer.microsoft.com/api/stac/v1",267 modifier=planetary_computer.sign_inplace,268)269270search = catalog.search(271 collections=["sentinel-2-l2a"],272 bbox=[-122.5, 37.7, -122.3, 37.9],273 datetime="2023-01-01/2023-12-31",274 query={"eo:cloud_cover": {"lt": 20}},275)276277# Load as xarray (cloud-native!)278data = odc.stac.load(279 list(search.get_items())[:5],280 bands=["B02", "B03", "B04", "B08"],281 crs="EPSG:32610",282 resolution=10,283)284285# Calculate NDVI on xarray286ndvi = (data.B08 - data.B04) / (data.B08 + data.B04)287```288289### Cloud-Optimized GeoTIFF (COG)290291```python292import rasterio293from rasterio.session import AWSSession294295# Read COG directly from cloud (partial reads)296session = AWSSession(aws_access_key_id=..., aws_secret_access_key=...)297with rasterio.open('s3://bucket/path.tif', session=session) as src:298 # Read only window of interest299 window = ((1000, 2000), (1000, 2000))300 subset = src.read(1, window=window)301302# Write COG303with rasterio.open('output.tif', 'w', **profile,304 tiled=True, blockxsize=256, blockysize=256,305 compress='DEFLATE', predictor=2) as dst:306 dst.write(data)307308# Validate COG309from rio_cogeo.cogeo import cog_validate310cog_validate('output.tif')311```312313## Performance Tips314315```python316# 1. Spatial indexing (10-100x faster queries)317gdf.sindex # Auto-created by GeoPandas318319# 2. Chunk large rasters320with rasterio.open('large.tif') as src:321 for i, window in src.block_windows(1):322 block = src.read(1, window=window)323324# 3. Dask for big data325import dask.array as da326dask_array = da.from_rasterio('large.tif', chunks=(1, 1024, 1024))327328# 4. Use Arrow for I/O329gdf.to_file('output.gpkg', use_arrow=True)330331# 5. GDAL caching332from osgeo import gdal333gdal.SetCacheMax(2**30) # 1GB cache334335# 6. Parallel processing336rf = RandomForestClassifier(n_jobs=-1) # All cores337```338339## Best Practices3403411. **Always check CRS** before spatial operations3422. **Use projected CRS** for area/distance calculations3433. **Validate geometries**: `gdf = gdf[gdf.is_valid]`3444. **Handle missing data**: `gdf['geometry'] = gdf['geometry'].fillna(None)`3455. **Use efficient formats**: GeoPackage > Shapefile, Parquet for large data3466. **Apply cloud masking** to optical imagery3477. **Preserve lineage** for reproducible research3488. **Use appropriate resolution** for your analysis scale349350## Detailed Documentation351352- **[Coordinate Systems](references/coordinate-systems.md)** - CRS fundamentals, UTM, transformations353- **[Core Libraries](references/core-libraries.md)** - GDAL, Rasterio, GeoPandas, Shapely354- **[Remote Sensing](references/remote-sensing.md)** - Satellite missions, spectral indices, SAR355- **[Machine Learning](references/machine-learning.md)** - Deep learning, CNNs, GNNs for RS356- **[GIS Software](references/gis-software.md)** - QGIS, ArcGIS, GRASS integration357- **[Scientific Domains](references/scientific-domains.md)** - Marine, hydrology, agriculture, forestry358- **[Advanced GIS](references/advanced-gis.md)** - 3D GIS, spatiotemporal, topology359- **[Big Data](references/big-data.md)** - Distributed processing, GPU acceleration360- **[Industry Applications](references/industry-applications.md)** - Urban planning, disaster management361- **[Programming Languages](references/programming-languages.md)** - Python, R, Julia, JS, C++, Java, Go, Rust362- **[Data Sources](references/data-sources.md)** - Satellite catalogs, APIs363- **[Troubleshooting](references/troubleshooting.md)** - Common issues, debugging, error reference364- **[Code Examples](references/code-examples.md)** - 500+ examples365366---367368**GeoMaster covers everything from basic GIS operations to advanced remote sensing and machine learning.**