Geohexgrid

Tests:

A tiny Python 3.10+ library for making geographic flat-top hexagonal grids like QGIS’s create grid function. That’s it. Not designed for making other kinds of grids or discrete global grid systems.

Here a hexagonal grid, or hex grid for short, is a finite subset of a hexagonal tiling. A hexagonal tiling is a covering of the plane with regular hexagons in which exactly three hexagons meet at each vertex. (For more details, see the Wikipedia article on hexagonal tilings.) The circumradius of a hex grid is the circumradius of any one of its hexagons, that is, the radius of a circle circumscribing any one of the hexagons. This library favors the word ‘grid’ over ‘tiling’, because ‘grid’ is used more often in geographic context, the main context of this library.

The two main features of this library are

• Making a flat-top hexagonal grid of given circumradius that minimally covers a GeoDataFrame of features, where distance units come from the GeoDataFrame’s coordinate reference system (CRS), e.g. no units for no CRS, metres for the New Zealand Transverse Mercator (NZTM) CRS, and decimal degrees for the WGS84 CRS.

• By default, hex grids made with a common CRS and circumradius share an origin and thus have equal hexagons (and hexagon IDs) where they overlap. In other words, the grids share a single (infinite) hexagonal tiling of the plane, which is useful when reconciling multiple grids across different geographic areas.

The main non-feature of this library is

• Making any other kind of grid, e.g. ones with pointy-top hexagons, squares, triangles, kisrhombilles, Penrose tiles…

Here’s a typical example.

import geopandas as gpd
import geohexgrid as ghg

# Load New Zealand territorial authorities projected in EPSG 2193 (NZTM)
nz = gpd.read_file(DATA_DIR / "nz_tas.gpkg")

# Cover it minimally with hexagons of circumradius 10 kilometres
grid = ghg.make_grid_from_gdf(nz, R=10_000)

# Plot
base = nz.plot(color="black", figsize=(10, 10), aspect="equal")
grid.plot(ax=base, color="white", edgecolor="red", alpha=0.5)

But why hexagons?! Because hexagons are the bestagons. More seriously, no one grid type works best for all geographic applications. MRCagney, this library’s funder, often works with isochrones, which favor simple convex equal area grids with equidistant neighbor cells, that is, hex grids.