Highway2Vec Embedder¶

In [1]:

Copied!

from IPython.display import display

from srai.plotting import plot_numeric_data, plot_regions
from IPython.display import display

from srai.plotting import plot_numeric_data, plot_regions

Get an area to embed¶

In [2]:

Copied!

from srai.regionalizers import geocode_to_region_gdf

area_gdf = geocode_to_region_gdf("Wrocław, PL")
plot_regions(area_gdf, tiles_style="CartoDB positron")
from srai.regionalizers import geocode_to_region_gdf

area_gdf = geocode_to_region_gdf("Wrocław, PL")
plot_regions(area_gdf, tiles_style="CartoDB positron")

Out[2]:

Make this Notebook Trusted to load map: File -> Trust Notebook

Regionalize the area with a regionalizer¶

In [3]:

Copied!





from srai.regionalizers import H3Regionalizer

regionalizer = H3Regionalizer(9)
regions_gdf = regionalizer.transform(area_gdf)
print(len(regions_gdf))
display(regions_gdf.head(3))
plot_regions(regions_gdf, tiles_style="CartoDB positron")
from srai.regionalizers import H3Regionalizer

regionalizer = H3Regionalizer(9)
regions_gdf = regionalizer.transform(area_gdf)
print(len(regions_gdf))
display(regions_gdf.head(3))
plot_regions(regions_gdf, tiles_style="CartoDB positron")

	geometry
region_id
891e2042dd7ffff	POLYGON ((17.01563 51.17702, 17.01567 51.17534...
891e2042b1bffff	POLYGON ((17.02152 51.13682, 17.02155 51.13514...
891e2050b0bffff	POLYGON ((16.91712 51.20177, 16.91716 51.20009...

Out[3]:

Make this Notebook Trusted to load map: File -> Trust Notebook

Download a road infrastructure for the area¶

In [4]:

Copied!

from srai.loaders import OSMNetworkType, OSMWayLoader

loader = OSMWayLoader(OSMNetworkType.DRIVE)
nodes_gdf, edges_gdf = loader.load(area_gdf)

display(nodes_gdf.head(3))
display(edges_gdf.head(3))

ax = edges_gdf.plot(linewidth=1, figsize=(12, 7))
nodes_gdf.plot(ax=ax, markersize=3, color="red")
from srai.loaders import OSMNetworkType, OSMWayLoader

loader = OSMWayLoader(OSMNetworkType.DRIVE)
nodes_gdf, edges_gdf = loader.load(area_gdf)

display(nodes_gdf.head(3))
display(edges_gdf.head(3))

ax = edges_gdf.plot(linewidth=1, figsize=(12, 7))
nodes_gdf.plot(ax=ax, markersize=3, color="red")

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/srai/loaders/osm_way_loader/osm_way_loader.py:229: FutureWarning: The clean_periphery argument has been deprecated and will be removed in the v2.0.0 release. Future behavior will be as though clean_periphery=True. See the OSMnx v2 migration guide: https://github.com/gboeing/osmnx/issues/1123
  G_directed = ox.graph_from_polygon(
/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/osmnx/_overpass.py:359: FutureWarning: `settings.timeout` is deprecated and will be removed in the v2.0.0 release: use `settings.requests_timeout` instead. See the OSMnx v2 migration guide: https://github.com/gboeing/osmnx/issues/1123
  overpass_settings = _make_overpass_settings()

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/osmnx/_overpass.py:369: FutureWarning: `settings.timeout` is deprecated and will be removed in the v2.0.0 release: use `settings.requests_timeout` instead. See the OSMnx v2 migration guide: https://github.com/gboeing/osmnx/issues/1123
  yield _overpass_request(data={"data": query_str})

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/srai/loaders/osm_way_loader/osm_way_loader.py:237: FutureWarning: The `get_undirected` function is deprecated and will be removed in the v2.0.0 release. Replace it with `convert.to_undirected` instead. See the OSMnx v2 migration guide: https://github.com/gboeing/osmnx/issues/1123
  G_undirected = ox.utils_graph.get_undirected(G_directed)

	y	x	street_count	highway	ref	geometry
osmid
95584835	51.083111	17.049513	4	NaN	NaN	POINT (17.04951 51.08311)
95584841	51.084699	17.064367	3	NaN	NaN	POINT (17.06437 51.08470)
95584850	51.083328	17.035057	4	NaN	NaN	POINT (17.03506 51.08333)

	oneway	lanes-1	lanes-2	lanes-3	lanes-4	lanes-5	lanes-6	lanes-7	lanes-8	lanes-9	...	bicycle-official	lit-yes	lit-no	lit-sunset-sunrise	lit-24/7	lit-automatic	lit-disused	lit-limited	lit-interval	geometry
feature_id
0	0	0	1	0	0	0	0	0	0	0	...	0	1	0	0	0	0	0	0	0	LINESTRING (17.04951 51.08311, 17.04947 51.083...
1	0	0	1	0	0	0	0	0	0	0	...	0	1	0	0	0	0	0	0	0	LINESTRING (17.04951 51.08311, 17.04933 51.083...
2	0	0	1	1	1	0	0	0	0	0	...	0	1	0	0	0	0	0	0	0	LINESTRING (17.05357 51.08301, 17.05335 51.082...

3 rows × 219 columns

Out[4]:

<Axes: >

No description has been provided for this image

Find out which edges correspond to which regions¶

In [5]:

Copied!

from srai.joiners import IntersectionJoiner

joiner = IntersectionJoiner()
joint_gdf = joiner.transform(regions_gdf, edges_gdf)
joint_gdf
from srai.joiners import IntersectionJoiner

joiner = IntersectionJoiner()
joint_gdf = joiner.transform(regions_gdf, edges_gdf)
joint_gdf

Out[5]:


region_id	feature_id
891e2042dd7ffff	6270
	6265
	6266
	6242
	6263
...	...
891e204720bffff	2694
	5285
	2696
	2695
	9047

15545 rows × 0 columns

Get the embeddings for regions based on the road infrastructure¶

In [6]:

Copied!

from pytorch_lightning import seed_everything

from srai.embedders import Highway2VecEmbedder

seed_everything(42)

embedder = Highway2VecEmbedder()
embeddings = embedder.fit_transform(regions_gdf, edges_gdf, joint_gdf)
embeddings
from pytorch_lightning import seed_everything

from srai.embedders import Highway2VecEmbedder

seed_everything(42)

embedder = Highway2VecEmbedder()
embeddings = embedder.fit_transform(regions_gdf, edges_gdf, joint_gdf)
embeddings

Seed set to 42

GPU available: False, used: False

TPU available: False, using: 0 TPU cores

IPU available: False, using: 0 IPUs

HPU available: False, using: 0 HPUs

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py:75: Starting from v1.9.0, `tensorboardX` has been removed as a dependency of the `pytorch_lightning` package, due to potential conflicts with other packages in the ML ecosystem. For this reason, `logger=True` will use `CSVLogger` as the default logger, unless the `tensorboard` or `tensorboardX` packages are found. Please `pip install lightning[extra]` or one of them to enable TensorBoard support by default

  | Name    | Type       | Params
---------------------------------------
0 | encoder | Sequential | 16.0 K
1 | decoder | Sequential | 16.2 K
---------------------------------------
32.1 K    Trainable params
0         Non-trainable params
32.1 K    Total params
0.128     Total estimated model params size (MB)

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/pytorch_lightning/trainer/connectors/data_connector.py:441: The 'train_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=3` in the `DataLoader` to improve performance.

`Trainer.fit` stopped: `max_epochs=10` reached.

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/srai/embedders/highway2vec/embedder.py:75: FutureWarning: The behavior of array concatenation with empty entries is deprecated. In a future version, this will no longer exclude empty items when determining the result dtype. To retain the old behavior, exclude the empty entries before the concat operation.
  embeddings_joint = joint_gdf.join(embeddings_df)

Out[6]:

	0	1	2	3	4	5	6	7	8	9	...	20	21	22	23	24	25	26	27	28	29
region_id
891e2040007ffff	-0.539541	-0.147202	0.231133	-0.155268	0.318578	-0.122770	0.081642	-0.649768	-0.241923	0.726528	...	0.307526	-0.507819	0.135545	0.039534	0.025653	0.079719	0.150593	0.348484	0.157026	0.348836
891e2040013ffff	-0.153741	-0.081649	0.326839	-0.272349	0.590248	-0.133988	-0.032478	-0.647943	-0.416105	0.836500	...	0.403983	-0.796961	0.215343	-0.348054	0.034720	0.147255	0.421888	0.416943	0.242064	0.397768
891e2040017ffff	-0.265606	-0.259180	0.262159	-0.227918	0.437086	-0.231525	0.185855	-0.838344	-0.175982	0.586379	...	0.246594	-0.570084	0.175362	-0.032558	0.046645	0.103537	0.481308	0.251798	0.145708	0.375634
891e2040023ffff	-0.257300	-0.354628	0.342268	-0.040691	0.502657	-0.160569	0.144855	-0.433879	-0.077589	0.586579	...	0.139541	-0.493104	0.179853	-0.014301	-0.046311	0.260333	0.386325	0.069908	-0.028361	0.441664
891e2040027ffff	-0.577240	-0.245056	0.116357	-0.331536	0.622523	0.034275	0.088231	-0.758699	-0.301155	0.729003	...	0.018124	-0.601429	0.183795	-0.155685	-0.012062	0.326336	0.176341	0.187382	0.018452	0.588956
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
891e2055bc3ffff	-0.531588	0.053953	0.218250	-0.149345	0.485677	-0.337040	0.185409	-0.816564	-0.293617	0.745303	...	0.466325	-0.575672	-0.017546	-0.210806	-0.246534	0.317339	0.094239	0.400965	0.386304	0.760467
891e2055bc7ffff	-0.312561	0.180587	0.216111	-0.196476	0.789139	-0.478202	0.298272	-0.609006	-0.374453	0.736907	...	0.351175	-0.600261	0.029559	-0.233984	-0.333623	0.331489	-0.026461	0.374098	0.311198	0.490092
891e2055bcbffff	-0.739363	0.035218	0.297716	-0.088070	0.444381	-0.183434	0.193246	-1.048380	-0.251863	0.719740	...	0.605018	-0.496426	-0.185179	-0.304660	-0.223717	0.455643	0.172639	0.451482	0.569794	1.042124
891e205a967ffff	-0.643766	0.128776	0.195003	-0.297602	0.324029	0.154858	-0.095509	-0.372894	-0.371210	0.532472	...	-0.283816	-0.305983	0.359435	0.000081	0.210609	0.025653	0.207406	0.221255	0.052564	0.575522
891e205a9a7ffff	-1.478251	-1.421816	0.651385	-1.063441	0.967914	-0.829468	0.323825	-0.825694	-0.475194	1.608099	...	0.584450	-0.549811	-0.137307	-0.018250	-0.286358	0.310460	0.385605	0.967372	0.244139	0.914667

2032 rows × 30 columns

In [7]:

Copied!

from sklearn.cluster import KMeans

clusterizer = KMeans(n_clusters=5, random_state=42)
clusterizer.fit(embeddings)

embeddings["cluster"] = clusterizer.labels_
embeddings
from sklearn.cluster import KMeans

clusterizer = KMeans(n_clusters=5, random_state=42)
clusterizer.fit(embeddings)

embeddings["cluster"] = clusterizer.labels_
embeddings

Out[7]:

	0	1	2	3	4	5	6	7	8	9	...	21	22	23	24	25	26	27	28	29	cluster
region_id
891e2040007ffff	-0.539541	-0.147202	0.231133	-0.155268	0.318578	-0.122770	0.081642	-0.649768	-0.241923	0.726528	...	-0.507819	0.135545	0.039534	0.025653	0.079719	0.150593	0.348484	0.157026	0.348836	3
891e2040013ffff	-0.153741	-0.081649	0.326839	-0.272349	0.590248	-0.133988	-0.032478	-0.647943	-0.416105	0.836500	...	-0.796961	0.215343	-0.348054	0.034720	0.147255	0.421888	0.416943	0.242064	0.397768	0
891e2040017ffff	-0.265606	-0.259180	0.262159	-0.227918	0.437086	-0.231525	0.185855	-0.838344	-0.175982	0.586379	...	-0.570084	0.175362	-0.032558	0.046645	0.103537	0.481308	0.251798	0.145708	0.375634	0
891e2040023ffff	-0.257300	-0.354628	0.342268	-0.040691	0.502657	-0.160569	0.144855	-0.433879	-0.077589	0.586579	...	-0.493104	0.179853	-0.014301	-0.046311	0.260333	0.386325	0.069908	-0.028361	0.441664	3
891e2040027ffff	-0.577240	-0.245056	0.116357	-0.331536	0.622523	0.034275	0.088231	-0.758699	-0.301155	0.729003	...	-0.601429	0.183795	-0.155685	-0.012062	0.326336	0.176341	0.187382	0.018452	0.588956	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
891e2055bc3ffff	-0.531588	0.053953	0.218250	-0.149345	0.485677	-0.337040	0.185409	-0.816564	-0.293617	0.745303	...	-0.575672	-0.017546	-0.210806	-0.246534	0.317339	0.094239	0.400965	0.386304	0.760467	4
891e2055bc7ffff	-0.312561	0.180587	0.216111	-0.196476	0.789139	-0.478202	0.298272	-0.609006	-0.374453	0.736907	...	-0.600261	0.029559	-0.233984	-0.333623	0.331489	-0.026461	0.374098	0.311198	0.490092	4
891e2055bcbffff	-0.739363	0.035218	0.297716	-0.088070	0.444381	-0.183434	0.193246	-1.048380	-0.251863	0.719740	...	-0.496426	-0.185179	-0.304660	-0.223717	0.455643	0.172639	0.451482	0.569794	1.042124	2
891e205a967ffff	-0.643766	0.128776	0.195003	-0.297602	0.324029	0.154858	-0.095509	-0.372894	-0.371210	0.532472	...	-0.305983	0.359435	0.000081	0.210609	0.025653	0.207406	0.221255	0.052564	0.575522	3
891e205a9a7ffff	-1.478251	-1.421816	0.651385	-1.063441	0.967914	-0.829468	0.323825	-0.825694	-0.475194	1.608099	...	-0.549811	-0.137307	-0.018250	-0.286358	0.310460	0.385605	0.967372	0.244139	0.914667	2

2032 rows × 31 columns

In [8]:

Copied!

plot_numeric_data(regions_gdf, "cluster", embeddings)
plot_numeric_data(regions_gdf, "cluster", embeddings)

Out[8]:

Make this Notebook Trusted to load map: File -> Trust Notebook