earthmover-colormaps 0.1.0

Perceptually uniform colormaps from the Earthmover brand palette

earthmover-colormaps

Perceptually uniform colormaps built from the Earthmover brand palette. Designed for scientific visualization — monotonic lightness, colorblind-safe, and zero required dependencies.

Install

uv add earthmover-colormaps

Or with pip:

pip install earthmover-colormaps

Or from source:

uv add git+https://github.com/earth-mover/earthmover-colormaps.git

Quick start

import earthmover_colormaps  # registers colormaps with matplotlib on import
import xarray as xr

ds = xr.tutorial.open_dataset("air_temperature")
ds.air.isel(time=0).plot(cmap="em.bloom")

import earthmover_colormaps registers everything with matplotlib's global registry, so the names work anywhere a cmap= argument is accepted (plt.imshow, plt.pcolormesh, xarray.plot(), cartopy, …).

Colormaps

Name	Type
em.violet	Sequential
em.lime	Sequential
em.signal	Sequential
em.diverging	Diverging
em.ocean	Sequential
em.bloom	Sequential
em.cycle	Cyclic

Every colormap has a reversed variant (append _r): "em.signal_r", "em.violet_r", etc.

Access patterns

import earthmover_colormaps

# 1. String name (after import registers them with matplotlib)
plt.imshow(data, cmap="em.signal")

# 2. Attribute access (short name, no "em." prefix)
earthmover_colormaps.signal
earthmover_colormaps.diverging
earthmover_colormaps.bloom

# 3. Dict access (full name)
earthmover_colormaps.cm["em.signal"]
earthmover_colormaps.cm["em.signal_r"]

Setting as default

To use an Earthmover colormap as your default across all plots:

import earthmover_colormaps
import matplotlib as mpl

mpl.rcParams["image.cmap"] = "em.signal"

Or in a matplotlibrc file:

image.cmap: em.signal

Design

Linear J' lightness in CAM02-UCS, gamut-clipped chroma, validated under simulated deuteranopia / protanopia / tritanopia.

To inspect any map (or compare against a matplotlib builtin):

uv run --group design python tools/compare_gui.py

Releasing

Versions come from git tags via hatch-vcs — no edits to pyproject.toml. To cut a release:

git tag v0.2.0
git push origin v0.2.0
gh release create v0.2.0 --generate-notes

Publishing the GitHub release fires .github/workflows/publish.yml, which builds and uploads to PyPI via OIDC trusted publishing.

License

Apache 2.0. See LICENSE.