Hex2vec embedder

Run this notebook in Google Colab:

Remember to install the srai library before running the notebook:

%pip install srai[all]

In [1]:

Copied!





import warnings

import matplotlib.pyplot as plt
import pandas as pd
from pytorch_lightning import seed_everything
from pytorch_lightning.loggers import CSVLogger
from srai.embedders import Hex2VecEmbedder
from srai.joiners import IntersectionJoiner
from srai.loaders import OSMOnlineLoader
from srai.neighbourhoods import H3Neighbourhood
from srai.plotting import plot_numeric_data, plot_regions
from srai.regionalizers import H3Regionalizer, geocode_to_region_gdf
import warnings

import matplotlib.pyplot as plt
import pandas as pd
from pytorch_lightning import seed_everything
from pytorch_lightning.loggers import CSVLogger
from srai.embedders import Hex2VecEmbedder
from srai.joiners import IntersectionJoiner
from srai.loaders import OSMOnlineLoader
from srai.neighbourhoods import H3Neighbourhood
from srai.plotting import plot_numeric_data, plot_regions
from srai.regionalizers import H3Regionalizer, geocode_to_region_gdf

In [2]:

Copied!

SEED = 71
seed_everything(SEED)
SEED = 71
seed_everything(SEED)

Seed set to 71

Out[2]:

Load data from OSM¶

First use geocoding to get the area

In [3]:

Copied!

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

Out[3]:

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

Next, download the data for the selected region and the specified tags. We're using OSMOnlineLoader here, as it's faster for low numbers of tags. In a real life scenario with more tags, you would likely want to use the OSMPbfLoader.

In [4]:

Copied!





tags = {
    "leisure": "park",
    "landuse": "forest",
    "amenity": ["bar", "restaurant", "cafe"],
    "water": "river",
    "sport": "soccer",
}
loader = OSMOnlineLoader()

features_gdf = loader.load(area_gdf, tags)

folium_map = plot_regions(
    area_gdf, colormap=["rgba(0,0,0,0)"], tiles_style="CartoDB positron"
)
features_gdf.explore(m=folium_map)
tags = {
    "leisure": "park",
    "landuse": "forest",
    "amenity": ["bar", "restaurant", "cafe"],
    "water": "river",
    "sport": "soccer",
}
loader = OSMOnlineLoader()

features_gdf = loader.load(area_gdf, tags)

folium_map = plot_regions(
    area_gdf, colormap=["rgba(0,0,0,0)"], tiles_style="CartoDB positron"
)
features_gdf.explore(m=folium_map)

Out[4]:

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

Prepare the data for embedding¶

After downloading the data, we need to prepare it for embedding. Namely - we need to regionalize the selected area, and join the features with regions.

In [5]:

Copied!

regionalizer = H3Regionalizer(resolution=9)
regions_gdf = regionalizer.transform(area_gdf)
plot_regions(regions_gdf, tiles_style="CartoDB positron")
regionalizer = H3Regionalizer(resolution=9)
regions_gdf = regionalizer.transform(area_gdf)
plot_regions(regions_gdf, tiles_style="CartoDB positron")

Out[5]:

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

In [6]:

Copied!

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

Out[6]:


region_id	feature_id
891e2040897ffff	node/280727473
891e2040d4bffff	node/300461026
891e2040d4bffff	node/300461036
891e2040d5bffff	node/300461042
891e2040887ffff	node/300461045
...	...
891e2042e73ffff	way/1427496434
891e2040a8fffff	way/1428809179
891e2045203ffff	way/1429016156
891e2045217ffff	way/1429016156
891e2040e43ffff	way/1429586876

4192 rows × 0 columns

Embedding¶

After preparing the data we can proceed with generating embeddings for the regions.

In [7]:

Copied!





neighbourhood = H3Neighbourhood(regions_gdf)
embedder = Hex2VecEmbedder([15, 10])

csv_logger = CSVLogger(save_dir="hex2vec_logs")

with warnings.catch_warnings():
    warnings.simplefilter("ignore")
    embeddings = embedder.fit_transform(
        regions_gdf,
        features_gdf,
        joint_gdf,
        neighbourhood,
        trainer_kwargs={"max_epochs": 5, "accelerator": "cpu", "logger": csv_logger},
        batch_size=100,
    )
embeddings
neighbourhood = H3Neighbourhood(regions_gdf)
embedder = Hex2VecEmbedder([15, 10])

csv_logger = CSVLogger(save_dir="hex2vec_logs")

with warnings.catch_warnings():
    warnings.simplefilter("ignore")
    embeddings = embedder.fit_transform(
        regions_gdf,
        features_gdf,
        joint_gdf,
        neighbourhood,
        trainer_kwargs={"max_epochs": 5, "accelerator": "cpu", "logger": csv_logger},
        batch_size=100,
    )
embeddings

GPU available: False, used: False

TPU available: False, using: 0 TPU cores

HPU available: False, using: 0 HPUs

  | Name    | Type       | Params | Mode 
-----------------------------------------------
0 | encoder | Sequential | 280    | train
-----------------------------------------------
280       Trainable params
0         Non-trainable params
280       Total params
0.001     Total estimated model params size (MB)
4         Modules in train mode
0         Modules in eval mode

`Trainer.fit` stopped: `max_epochs=5` reached.

Out[7]:

	0	1	2	3	4	5	6	7	8	9
region_id
891e2041943ffff	-0.540500	0.368682	-0.502314	-0.796841	-0.066350	0.231252	-0.294797	0.403219	-0.200558	-0.282214
891e2041973ffff	0.041568	0.127272	0.433192	0.560574	-0.128483	0.006710	0.322844	-0.352005	0.398866	0.111660
891e204e633ffff	0.312323	-0.231362	-0.038064	0.301079	0.230049	-0.098567	-0.285888	-0.072734	-0.201323	0.377655
891e2051b07ffff	-0.540500	0.368682	-0.502314	-0.796841	-0.066350	0.231252	-0.294797	0.403219	-0.200558	-0.282214
891e2041873ffff	-0.275702	0.036633	-0.284333	-0.419740	0.028889	0.215045	-0.426803	0.316230	-0.118924	0.039350
...	...	...	...	...	...	...	...	...	...	...
891e2042913ffff	0.312323	-0.231362	-0.038064	0.301079	0.230049	-0.098567	-0.285888	-0.072734	-0.201323	0.377655
891e2042dbbffff	0.095703	-0.365364	0.175965	-0.416457	-0.229633	0.114342	0.170426	0.129182	-0.026517	-0.460903
891e20426dbffff	-0.275702	0.036633	-0.284333	-0.419740	0.028889	0.215045	-0.426803	0.316230	-0.118924	0.039350
891e204556fffff	-0.414475	0.007036	-0.279861	-0.851557	-0.149023	0.274329	-0.127790	0.443625	-0.177185	-0.404225
891e2045413ffff	0.041568	0.127272	0.433192	0.560574	-0.128483	0.006710	0.322844	-0.352005	0.398866	0.111660

3168 rows × 10 columns

In [8]:

Copied!





metrics_df = pd.read_csv(csv_logger.log_dir + "/metrics.csv").dropna(
    subset="train_f1_epoch"
)

fig, ax1 = plt.subplots(1, 1, figsize=(10, 5))
ax2 = ax1.twinx()

line1 = ax1.plot(metrics_df["epoch"], metrics_df["train_f1_epoch"])
line2 = ax2.plot(metrics_df["epoch"], metrics_df["train_loss_epoch"], color="orange")
ax1.legend(line1 + line2, ["F1", "Loss"], loc=7)
ax1.set_title("Training metrics")
ax1.set_ylabel("F1")
ax2.set_ylabel("Loss")
ax1.set_xlabel("Training epoch")
plt.show()
metrics_df = pd.read_csv(csv_logger.log_dir + "/metrics.csv").dropna(
    subset="train_f1_epoch"
)

fig, ax1 = plt.subplots(1, 1, figsize=(10, 5))
ax2 = ax1.twinx()

line1 = ax1.plot(metrics_df["epoch"], metrics_df["train_f1_epoch"])
line2 = ax2.plot(metrics_df["epoch"], metrics_df["train_loss_epoch"], color="orange")
ax1.legend(line1 + line2, ["F1", "Loss"], loc=7)
ax1.set_title("Training metrics")
ax1.set_ylabel("F1")
ax2.set_ylabel("Loss")
ax1.set_xlabel("Training epoch")
plt.show()

No description has been provided for this image

Visualizing the embeddings' similarity¶

In [9]:

Copied!

from sklearn.cluster import KMeans

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

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

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

embeddings["cluster"] = clusterizer.labels_
embeddings

Out[9]:

	0	1	2	3	4	5	6	7	8	9	cluster
region_id
891e2041943ffff	-0.540500	0.368682	-0.502314	-0.796841	-0.066350	0.231252	-0.294797	0.403219	-0.200558	-0.282214	4
891e2041973ffff	0.041568	0.127272	0.433192	0.560574	-0.128483	0.006710	0.322844	-0.352005	0.398866	0.111660	3
891e204e633ffff	0.312323	-0.231362	-0.038064	0.301079	0.230049	-0.098567	-0.285888	-0.072734	-0.201323	0.377655	0
891e2051b07ffff	-0.540500	0.368682	-0.502314	-0.796841	-0.066350	0.231252	-0.294797	0.403219	-0.200558	-0.282214	4
891e2041873ffff	-0.275702	0.036633	-0.284333	-0.419740	0.028889	0.215045	-0.426803	0.316230	-0.118924	0.039350	4
...	...	...	...	...	...	...	...	...	...	...	...
891e2042913ffff	0.312323	-0.231362	-0.038064	0.301079	0.230049	-0.098567	-0.285888	-0.072734	-0.201323	0.377655	0
891e2042dbbffff	0.095703	-0.365364	0.175965	-0.416457	-0.229633	0.114342	0.170426	0.129182	-0.026517	-0.460903	1
891e20426dbffff	-0.275702	0.036633	-0.284333	-0.419740	0.028889	0.215045	-0.426803	0.316230	-0.118924	0.039350	4
891e204556fffff	-0.414475	0.007036	-0.279861	-0.851557	-0.149023	0.274329	-0.127790	0.443625	-0.177185	-0.404225	4
891e2045413ffff	0.041568	0.127272	0.433192	0.560574	-0.128483	0.006710	0.322844	-0.352005	0.398866	0.111660	3

3168 rows × 11 columns

In [10]:

Copied!

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

Out[10]:

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