Environmental Feature Maps

The scripts/plot_environmental.py script visualizes the raw environmental data sampled onto the H3 grid by utils/geoutils.py. It produces publication-quality PNG maps for elevation, temperature, precipitation, land cover, and other features — useful for sanity-checking the input data before training.

Usage

# Plot standard environmental columns
python scripts/plot_environmental.py --input data/global_350km_ee.parquet

# Restrict to a geographic region and specific columns
python scripts/plot_environmental.py --input data/global_350km_ee.parquet \
    --bounds europe --columns elevation,temperature_mean

CLI Reference

Flag	Default	Description
--input / -i	required	Input GeoParquet file (from geoutils.py)
--outdir / -o	outputs/plots	Output directory for PNGs
--sample-limit	200000	Max cells to plot (random sample). Use None or -1 for no limit
--bounds	—	Geographic bounds: named region or 4 floats (west south east north)
--columns	standard set	Comma-separated list of columns to plot

Output

One PNG per environmental column (e.g. elevation.png, temperature_mean.png). Land-cover maps use a categorical color palette with a legend; continuous variables use sequential color maps.