Voronoi diagram for polygons is a tool to create a Voronoi diagram also known as Thiessen polygons for polygons. It's based on Shapely and GeoPandas. There are lots of tools to create a Voronoi diagram for points, for example Create Thiessen Polygons (Analysis) in ArcGIS Pro or ArcGIS Desktop, Voronoi Polygons in QGIS, or voronoi_diagram in Shapely. All of them are really cool. How about a Voronoi diagram for polygons? That's what this tool does.

Important! You have to install or upgrade to the latest developing version of Shapely before install Voronoi diagram for polygons

• Installation, update and uninstallation
  • Dependencies
  • To install
  • To update
  • To uninstall
• Usage
• Known shortages
• How to cite?
• Changelog

Important! It's based on voronoi_diagram from Shapely which is new in version 1.8.dev0. As of today, it is still a developing version. (2020-07-26) You have to install or upgrade to the latest developing version from source firstly:

pip install git+https://github.com/Toblerity/Shapely

pip install --upgrade git+https://github.com/Toblerity/Shapely

Quick installation with pip:

pip install voronoi-diagram-for-polygons

Or from github:

pip install git+https://github.com/longavailable/voronoi-diagram-for-polygons

Also, you can just copy related functions from longsgis/longsgis.py to your work.

pip install --upgrade voronoi-diagram-for-polygons

pip uninstall voronoi-diagram-for-polygons

See tests/01voronoiDiagram4plg.py.

import geopandas as gpd
from longsgis import voronoiDiagram4plg

builtup = gpd.read_file('input.geojson'); builtup.crs = 32650
boundary = gpd.read_file('boundary.geojson'); boundary.crs = 32650
vd = voronoiDiagram4plg(builtup, boundary)
vd.to_file('output.geojson', driver='GeoJSON')

• It may produce multipolygons (consisted by some unconnected polygons) around the boundary.

  

• Special input may cause overlap. See the following:

  

  To avoid this, I recommend reasonable preprocessing of the input, but use a buffer operation with high-resolution carefully. A buffer operation with high-resolution will result in circular arcs, which will generate too many vertices in a local area. This may trigger other bugs. In my practices, the following code snippet worked well.

def bufferDissolve(gdf, distance, join_style=3):	
	'''Create buffer and dissolve thoese intersects.
	
	Parameters:
		gdf: 
			Type: geopandas.GeoDataFrame
		distance: radius of the buffer
			Type: float
	Returns:
		gdf_bf: buffered and dissolved GeoDataFrame
			Type: geopandas.GeoDataFrame
	'''
	#create buffer and dissolve by invoking `unary_union`
	smp = gdf.buffer(distance, join_style).unary_union
	#convert to GeoSeries and explode to single polygons
	gs = gpd.GeoSeries([smp]).explode()
	#convert to GeoDataFrame
	gdf_bf = gpd.GeoDataFrame(geometry=gs, crs=gdf.crs).reset_index(drop=True)
	return gdf_bf

If this tool is useful to your research, star and cite it as below:

Xiaolong Liu, & Meixiu Yu. (2020, July 26). longavailable/voronoi-diagram-for-polygons. Zenodo. 
http://doi.org/10.5281/zenodo.3960407

Easily, you can import it to Mendeley.

• First release.