Proximity bands¶

Proximity bands are street-based morphological elements following the pedestrian perspective of urban fabric analysis described by Araldi and Fusco [2019]. Instead of representing the city from parcels, blocks, or buildings, they describe the portion of space assigned to each street segment within a fixed walking-side distance.

In momepy, momepy.proximity_bands builds these elements as a Voronoi partition of the area within a buffer around street geometry. The result is useful when a character should be measured from the street, for example the built form or land cover that can be associated with a pedestrian moving along a street segment.

The distance is expressed in the units of the CRS, so the input should use a projected CRS.

import geopandas as gpd
import mapclassify
import momepy

path = momepy.datasets.get_path("bubenec")
streets = gpd.read_file(path, layer="streets")

We start with street centerlines. The Bubenec sample data are already projected in meters.

ax = streets.plot(color="black", linewidth=2, figsize=(8, 8))
ax.set_axis_off()

A standard proximity band assigns the area within band distance to the nearest street segment. Here, each segment gets one polygon, clipped to a 20 metre buffer around the street network.

bands = momepy.proximity_bands(streets, band=20)
bands.head()

unique_color = mapclassify.greedy(bands)
ax = bands.plot(
    unique_color,
    categorical=True,
    edgecolor="white",
    linewidth=0.4,
    figsize=(8, 8),
)
streets.plot(ax=ax, color="black", linewidth=1)
ax.set_axis_off()

The default output has the same index as the input streets, which makes it straightforward to attach measurements back to street segments.

bands.index.equals(streets.index)

True

Single-sided bands¶

Araldi and Fusco’s pedestrian perspective treats the street as the observation axis. In many situations, the two sides of a street expose different urban fabric: buildings, plots, or open space may differ across the centerline. Passing single_sided=True splits the band along the street geometry and returns separate polygons for each side.

single_sided = momepy.proximity_bands(streets, band=20, single_sided=True)
single_sided.head()

unique_color = mapclassify.greedy(single_sided)
ax = single_sided.plot(
    unique_color,
    categorical=True,
    edgecolor="white",
    linewidth=0.4,
    figsize=(8, 8),
)
streets.plot(ax=ax, color="black", linewidth=1)
ax.set_axis_off()

Use the two-sided version when every street segment should be represented by one element. Use the single-sided version when the analysis needs to distinguish the left and right sides of the pedestrian environment.