As explained in the Data Structure chapter,
momepy relies on links between different morphological elements. Each element needs ID, and each of the small-scale elements also needs to know the ID of the relevant higher-scale element. The case of block ID is explained in the previous chapter,
momepy.Blocks generates it together with blocks gdf.
This notebook will explore how to link street network, both nodes and edges, to buildings and tessellation.
For linking street network edges to buildings (or tessellation or other elements),
momepy.get_network_id. It simply returns a
Series of network IDs for analysed gdf.
import momepy import geopandas as gpd import matplotlib.pyplot as plt
For illustration, we can use
bubenec dataset embedded in
path = momepy.datasets.get_path('bubenec') buildings = gpd.read_file(path, layer='buildings') streets = gpd.read_file(path, layer='streets') tessellation = gpd.read_file(path, layer='tessellation')
First, we have to be sure that streets segments have their unique IDs.
streets['nID'] = momepy.unique_id(streets)
Then we can link it to buildings. The only argument we might want to look at is
min_size, which should be a value such that if you build a box centred in each building centroid with edges of size
2 * min_size, you know a priori that at least one segment is intersected with the box. You can see it as a sort of tolerance.
buildings['nID'] = momepy.get_network_id(buildings, streets, 'nID', min_size=100)
Snapping: 100%|██████████| 144/144 [00:00<00:00, 1136.30it/s]
Generating centroids... Generating rtree...
f, ax = plt.subplots(figsize=(10, 10)) buildings.plot(ax=ax, column='nID', categorical=True, cmap='tab20b') streets.plot(ax=ax) ax.set_axis_off() plt.show()
Note: colormap does not have enough colours, that is why everything on the top-left looks the same. It is not.
The situation with nodes is slightly more complicated as you usually don't have or even need nodes. However,
momepy includes some functions which are calculated on nodes (mostly in
graph module). For that reason, we will pretend that we follow the usual workflow:
- Street network
- Street network - edges and nodes as separate
graph = momepy.gdf_to_nx(streets)
nodes, edges = momepy.nx_to_gdf(graph)
f, ax = plt.subplots(figsize=(10, 10)) edges.plot(ax=ax, column='nID', categorical=True, cmap='tab20b') nodes.plot(ax=ax, zorder=2) ax.set_axis_off() plt.show()
For attaching node ID to buildings, we will need both, nodes and edges. We have already determined which edge building belongs to, so now we only have to find out which end of the edge is the closer one. Nodes come from
momepy.nx_to_gdf automatically with node ID:
|0||1||POINT (1603585.640 6464428.774)|
|1||2||POINT (1603413.206 6464228.730)|
|2||3||POINT (1603268.502 6464060.781)|
|3||4||POINT (1603363.558 6464031.885)|
|4||5||POINT (1603607.303 6464181.853)|
The same ID is now inlcuded in edges as well, denoting each end of edge. (Length of the edge is also present as it was necessary to keep as an attribute for the graph.)
|0||LINESTRING (1603585.640 6464428.774, 1603413.2...||0||264.103950||1||2|
|1||LINESTRING (1603561.740 6464494.467, 1603564.6...||14||70.020202||1||9|
|2||LINESTRING (1603585.640 6464428.774, 1603603.0...||15||88.924305||1||7|
|3||LINESTRING (1603607.303 6464181.853, 1603592.8...||2||199.746503||2||5|
|4||LINESTRING (1603363.558 6464031.885, 1603376.5...||5||203.014090||2||4|
buildings['nodeID'] = momepy.get_node_id(buildings, nodes, edges, 'nodeID', 'nID')
100%|██████████| 144/144 [00:00<00:00, 226.13it/s]
f, ax = plt.subplots(figsize=(10, 10)) buildings.plot(ax=ax, column='nodeID', categorical=True, cmap='tab20b') nodes.plot(ax=ax, zorder=2) edges.plot(ax=ax, zorder=1) ax.set_axis_off() plt.show()
Index(['uID', 'geometry', 'nID', 'nodeID'], dtype='object')
columns = ['uID', 'nID', 'nodeID'] tessellation = tessellation.merge(buildings[columns], on='uID') tessellation.head()
|0||1.0||POLYGON ((1603578.489 6464344.527, 1603577.040...||0||1|
|1||2.0||POLYGON ((1603067.112 6464177.926, 1603054.848...||33||10|
|2||3.0||POLYGON ((1602978.618 6464156.859, 1603006.384...||10||12|
|3||4.0||POLYGON ((1603056.595 6464093.903, 1603011.539...||10||12|
|4||5.0||POLYGON ((1603110.459 6464114.367, 1603109.099...||8||12|
Now we should be able to link all elements together as needed for all types of morphometric analysis in