Porto taxi

Porto Taxi Dataset¶

The dataset covers one year of taxi trajectory data from 442 vehicles operating in Porto, Portugal, between July 1, 2013, and June 30, 2014. Each completed trip is categorized as either (A) taxi central–based, (B) stand–based, or (C) non–taxi central–based, reflecting whether the ride was initiated through the dispatch central, a taxi stand, or a random street pickup. Each record includes metadata such as trip origin type, taxi and call identifiers, day type, and a GPS trajectory encoded as a sequence of geographic coordinates.

Run this notebook in Google Colab:

Remember to install the srai library before running the notebook:

%pip install srai[all]

In [1]:

Copied!





import folium
import h3
from IPython.display import display
from srai.datasets import PortoTaxiDataset
import folium
import h3
from IPython.display import display
from srai.datasets import PortoTaxiDataset

In [2]:

Copied!

porto_taxi = PortoTaxiDataset()
porto_taxi = PortoTaxiDataset()

In [3]:

Copied!

type(porto_taxi.train_gdf), type(porto_taxi.test_gdf)
type(porto_taxi.train_gdf), type(porto_taxi.test_gdf)

Out[3]:

(NoneType, NoneType)

Get data using .load() method -> Default config (Travel time estimation)

In [4]:

Copied!

ds = porto_taxi.load()
ds.keys()
ds = porto_taxi.load()
ds.keys()

Out[4]:

dict_keys(['train', 'test'])

In [5]:

Copied!

type(porto_taxi.train_gdf), type(porto_taxi.test_gdf)
type(porto_taxi.train_gdf), type(porto_taxi.test_gdf)

Out[5]:

(geopandas.geodataframe.GeoDataFrame, geopandas.geodataframe.GeoDataFrame)

In [6]:

Copied!

ds["train"].head()
ds["train"].head()

Out[6]:

	trip_id	h3_sequence	avg_speed_per_hex	timestamp	day_type	call_type	taxi_id	geometry	duration
0	1372636853620000380	[8939220f383ffff, 8939220f393ffff, 8939220f067...	[1.927607771705634, 1.927607771705634, 1.92760...	[2013-07-01T00:00:53.000000, 2013-07-01T00:00:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A]	[C, C, C, C, C, C, C, C, C, C, C, C, C, C]	[20000380, 20000380, 20000380, 20000380, 20000...	LINESTRING (-8.61029 41.14075, -8.61031 41.140...	360.0
1	1372636858620000589	[8939220f063ffff, 8939220f073ffff, 8939220f00b...	[3.0560328387467504, 3.0560328387467504, 49.10...	[2013-07-01T00:00:58.000000, 2013-07-01T00:00:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A]	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C]	[20000589, 20000589, 20000589, 20000589, 20000...	LINESTRING (-8.61864 41.14141, -8.6185 41.1413...	330.0
2	1372636875620000233	[8939220f00fffff, 8939220f007ffff, 8939220f017...	[9.213600054030238, 9.213600054030238, 23.0525...	[2013-07-01T00:01:15.000000, 2013-07-01T00:01:...	[A, A, A, A, A, A, A, A, A, A, A]	[C, C, C, C, C, C, C, C, C, C, C]	[20000233, 20000233, 20000233, 20000233, 20000...	LINESTRING (-8.61989 41.14801, -8.62016 41.147...	315.0
3	1372636956620000167	[8939220f0afffff, 8939220f1dbffff, 8939220f0af...	[7.325408241610683, 7.325408241610683, 23.4882...	[2013-07-01T00:02:36.000000, 2013-07-01T00:02:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000167, 20000167, 20000167, 20000167, 20000...	LINESTRING (-8.6133 41.15439, -8.61326 41.1538...	465.0
4	1372636965620000231	[8939220f067ffff, 8939220f39bffff, 8939220f383...	[24.42802472301524, 24.42802472301524, 24.4280...	[2013-07-01T00:02:45.000000, 2013-07-01T00:02:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000231, 20000231, 20000231, 20000231, 20000...	LINESTRING (-8.6155 41.14067, -8.61335 41.1415...	375.0

In [7]:

Copied!

ds["test"].head()
ds["test"].head()

Out[7]:

	trip_id	h3_sequence	avg_speed_per_hex	timestamp	day_type	call_type	taxi_id	geometry	duration
0	1372638160620000456	[8939220f39bffff, 8939220f067ffff, 8939220f02b...	[24.38591002388346, 24.38591002388346, 24.3859...	[2013-07-01T00:22:40.000000, 2013-07-01T00:22:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000456, 20000456, 20000456, 20000456, 20000...	LINESTRING (-8.61215 41.14057, -8.61455 41.140...	435.0
1	1372638504620000600	[8939220c493ffff, 8939220e22fffff, 8939220e223...	[10.331445317923839, 10.331445317923839, 35.99...	[2013-07-01T00:28:24.000000, 2013-07-01T00:28:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[20000600, 20000600, 20000600, 20000600, 20000...	LINESTRING (-8.57362 41.17077, -8.57406 41.170...	945.0
2	1372639848620000424	[8939220f477ffff, 8939220f463ffff, 8939220f46b...	[3.3926150007803417, 3.3926150007803417, 13.70...	[2013-07-01T00:50:48.000000, 2013-07-01T00:50:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000424, 20000424, 20000424, 20000424, 20000...	LINESTRING (-8.63984 41.1598, -8.64034 41.1598...	600.0
3	1372639875620000009	[8939220e66fffff, 8939220e2d7ffff, 8939220e2c3...	[19.742960983412413, 19.742960983412413, 19.74...	[2013-07-01T00:51:15.000000, 2013-07-01T00:51:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000009, 20000009, 20000009, 20000009, 20000...	LINESTRING (-8.60007 41.1827, -8.5982 41.18226...	660.0
4	1372641721620000686	[8939220f02bffff, 8939220f02fffff, 8939220f027...	[12.46393351492334, 12.46393351492334, 12.4639...	[2013-07-01T01:22:01.000000, 2013-07-01T01:22:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[B, B, B, B, B, B, B, B, B, B, B, B, B, B, B, ...	[20000686, 20000686, 20000686, 20000686, 20000...	LINESTRING (-8.61253 41.14594, -8.61068 41.146...	810.0

Getting h3 trajectories with target valeus

In [8]:

Copied!

porto_taxi.resolution
porto_taxi.resolution

Out[8]:

In [9]:

Copied!

train_h3, _, test_h3 = porto_taxi.get_h3_with_labels()
train_h3, _, test_h3 = porto_taxi.get_h3_with_labels()

In [10]:

Copied!

train_h3
train_h3

Out[10]:

	trip_id	h3_sequence	duration
0	1372636853620000380	[8939220f383ffff, 8939220f393ffff, 8939220f067...	360.0
1	1372636858620000589	[8939220f063ffff, 8939220f073ffff, 8939220f00b...	330.0
2	1372636875620000233	[8939220f00fffff, 8939220f007ffff, 8939220f017...	315.0
3	1372636956620000167	[8939220f0afffff, 8939220f1dbffff, 8939220f0af...	465.0
4	1372636965620000231	[8939220f067ffff, 8939220f39bffff, 8939220f383...	375.0
...	...	...	...
358953	1404171287620000013	[8939220e66fffff, 8939220e2d7ffff, 8939220e2c3...	1485.0
358954	1404171347620000328	[8939220f1c7ffff, 8939220f1c3ffff, 8939220f1cf...	585.0
358955	1404171590620000307	[8939220f397ffff, 8939220f14bffff, 8939220f14f...	1185.0
358956	1404171854620000114	[8939220f477ffff, 8939220f473ffff, 8939220f477...	555.0
358957	1404172369620000171	[8939220f03bffff, 8939220f077ffff, 8939220f02b...	345.0

358958 rows × 3 columns

In [11]:

Copied!

test_h3
test_h3

Out[11]:

	trip_id	h3_sequence	duration
0	1372638160620000456	[8939220f39bffff, 8939220f067ffff, 8939220f02b...	435.0
1	1372638504620000600	[8939220c493ffff, 8939220e22fffff, 8939220e223...	945.0
2	1372639848620000424	[8939220f477ffff, 8939220f463ffff, 8939220f46b...	600.0
3	1372639875620000009	[8939220e66fffff, 8939220e2d7ffff, 8939220e2c3...	660.0
4	1372641721620000686	[8939220f02bffff, 8939220f02fffff, 8939220f027...	810.0
...	...	...	...
89735	1404170689620000387	[8939220f083ffff, 8939220f09bffff, 8939220f093...	450.0
89736	1404170875620000525	[8939220f507ffff, 8939220f517ffff, 8939220f513...	600.0
89737	1404171367620000670	[8939220f383ffff, 8939220f39bffff, 8939220f067...	435.0
89738	1404171374620000372	[8939220f6b3ffff, 8939220f6b7ffff, 8939220f46b...	555.0
89739	1404171442620000497	[8939220f08bffff, 8939220f08fffff, 8939220f0bb...	315.0

89740 rows × 3 columns

Get Human Mobility Prediction (HMP) data

In [12]:

Copied!

ds = porto_taxi.load(version="HMP")
ds = porto_taxi.load(version="HMP")

In [13]:

Copied!

ds["train"].head()
ds["train"].head()

Out[13]:

	trip_id	avg_speed_per_hex	timestamp	day_type	call_type	taxi_id	geometry	h3_sequence_x	h3_sequence_y
0	1372636875620000233	[9.213600054030238, 9.213600054030238, 23.0525...	[2013-07-01T00:01:15.000000, 2013-07-01T00:01:...	[A, A, A, A, A, A, A, A, A, A, A]	[C, C, C, C, C, C, C, C, C, C, C]	[20000233, 20000233, 20000233, 20000233, 20000...	LINESTRING (-8.61989 41.14801, -8.62016 41.147...	[8939220f00fffff, 8939220f007ffff, 8939220f017...	[8939220f1d3ffff, 8939220f1d3ffff]
1	1372636965620000231	[24.42802472301524, 24.42802472301524, 24.4280...	[2013-07-01T00:02:45.000000, 2013-07-01T00:02:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000231, 20000231, 20000231, 20000231, 20000...	LINESTRING (-8.6155 41.14067, -8.61335 41.1415...	[8939220f067ffff, 8939220f39bffff, 8939220f383...	[8939220c4cbffff, 8939220c4c3ffff, 8939220c4c3...
2	1372637084620000285	[9.733088667826689, 9.733088667826689, 9.73308...	[2013-07-01T00:04:44.000000, 2013-07-01T00:04:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000285, 20000285, 20000285, 20000285, 20000...	LINESTRING (-8.58568 41.14858, -8.58577 41.148...	[8939220f167ffff, 8939220f12bffff, 8939220f167...	[8939220c633ffff, 8939220c633ffff]
3	1372637247620000067	[6.5659677388795235, 6.5659677388795235, 34.91...	[2013-07-01T00:07:27.000000, 2013-07-01T00:07:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000067, 20000067, 20000067, 20000067, 20000...	LINESTRING (-8.62886 41.16097, -8.62853 41.160...	[8939220f097ffff, 8939220f083ffff, 8939220f087...	[8939220f023ffff, 8939220f023ffff]
4	1372637610620000497	[20.531007632411953, 20.531007632411953, 20.53...	[2013-07-01T00:13:30.000000, 2013-07-01T00:13:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[B, B, B, B, B, B, B, B, B, B, B, B, B, B, B, ...	[20000497, 20000497, 20000497, 20000497, 20000...	LINESTRING (-8.58514 41.16486, -8.58415 41.164...	[8939220e26bffff, 8939220f1b7ffff, 8939220f1b3...	[8939220f42bffff, 8939220f42fffff, 8939220f093...

Creating your own train_test split based on trajectory duration (version TTE) or length version (HMP).

Downloading version all without passing resolution, will return trajectories as linestring geometries.

In [14]:

Copied!

ds = porto_taxi.load(version="all")
ds.keys()
ds = porto_taxi.load(version="all")
ds.keys()

Out[14]:

dict_keys(['train'])

In [15]:

Copied!

ds["train"].head()
ds["train"].head()

Out[15]:

	trip_id	timestamp	call_type	origin_call	origin_stand	taxi_id	day_type	speed	geometry
0	1372636853620000380	[2013-07-01T00:00:53.000000, 2013-07-01T00:01:...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C]	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[20000380, 20000380, 20000380, 20000380, 20000...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A]	[1.927607771705634, 1.927607771705634, 43.9689...	LINESTRING (-8.61029 41.14075, -8.61031 41.140...
1	1372636858620000589	[2013-07-01T00:00:58.000000, 2013-07-01T00:01:...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C]	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[20000589, 20000589, 20000589, 20000589, 20000...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A]	[3.0560328387467504, 3.0560328387467504, 49.10...	LINESTRING (-8.61864 41.14141, -8.6185 41.1413...
2	1372636875620000233	[2013-07-01T00:01:15.000000, 2013-07-01T00:01:...	[C, C, C, C, C, C, C, C, C, C, C]	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[20000233, 20000233, 20000233, 20000233, 20000...	[A, A, A, A, A, A, A, A, A, A, A]	[9.213600054030238, 9.213600054030238, 23.0525...	LINESTRING (-8.61989 41.14801, -8.62016 41.147...
3	1372636956620000167	[2013-07-01T00:02:36.000000, 2013-07-01T00:03:...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[20000167, 20000167, 20000167, 20000167, 20000...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[7.325408241610683, 7.325408241610683, 23.4882...	LINESTRING (-8.6133 41.15439, -8.61326 41.1538...
4	1372636965620000231	[2013-07-01T00:02:45.000000, 2013-07-01T00:03:...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[nan, nan, nan, nan, nan, nan, nan, nan, nan, ...	[20000231, 20000231, 20000231, 20000231, 20000...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[24.42802472301524, 24.42802472301524, 70.2825...	LINESTRING (-8.6155 41.14067, -8.61335 41.1415...

Passing resolution parameter is neccessary for generation of trajectory in h3 style.

Resolution parameter is required to create h3 sequences from the linestring geometry.

In [16]:

Copied!

ds = porto_taxi.load(version="all", resolution=10)
ds.keys()
ds = porto_taxi.load(version="all", resolution=10)
ds.keys()

Out[16]:

dict_keys(['train'])

In [17]:

Copied!

ds["train"].head()
ds["train"].head()

Out[17]:

	trip_id	h3_sequence	avg_speed_per_hex	timestamp	day_type	call_type	taxi_id	geometry
0	1372636853620000380	[8a39220f391ffff, 8a39220f3917fff, 8a39220f064...	[1.927607771705634, 1.927607771705634, 1.92760...	[2013-07-01T00:00:53.000000, 2013-07-01T00:00:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A]	[C, C, C, C, C, C, C, C, C, C, C, C, C, C]	[20000380, 20000380, 20000380, 20000380, 20000...	LINESTRING (-8.61029 41.14075, -8.61031 41.140...
1	1372636858620000589	[8a39220f062ffff, 8a39220f0627fff, 8a39220f070...	[3.0560328387467504, 3.0560328387467504, 3.056...	[2013-07-01T00:00:58.000000, 2013-07-01T00:00:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A]	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C]	[20000589, 20000589, 20000589, 20000589, 20000...	LINESTRING (-8.61864 41.14141, -8.6185 41.1413...
2	1372636875620000233	[8a39220f00effff, 8a39220f0057fff, 8a39220f007...	[9.213600054030238, 9.213600054030238, 23.0525...	[2013-07-01T00:01:15.000000, 2013-07-01T00:01:...	[A, A, A, A, A, A, A, A, A, A, A]	[C, C, C, C, C, C, C, C, C, C, C]	[20000233, 20000233, 20000233, 20000233, 20000...	LINESTRING (-8.61989 41.14801, -8.62016 41.147...
3	1372636956620000167	[8a39220f0aeffff, 8a39220f0acffff, 8a39220f1d9...	[7.325408241610683, 7.325408241610683, 7.32540...	[2013-07-01T00:02:36.000000, 2013-07-01T00:02:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000167, 20000167, 20000167, 20000167, 20000...	LINESTRING (-8.6133 41.15439, -8.61326 41.1538...
4	1372636965620000231	[8a39220f0657fff, 8a39220f0647fff, 8a39220f064...	[24.42802472301524, 24.42802472301524, 24.4280...	[2013-07-01T00:02:45.000000, 2013-07-01T00:02:...	[A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, ...	[C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, ...	[20000231, 20000231, 20000231, 20000231, 20000...	LINESTRING (-8.6155 41.14067, -8.61335 41.1415...

In [18]:

Copied!

train, test = porto_taxi.train_test_split(
    target_column="trip_id", task="HMP", test_size=0.2, n_bins=7
)
train, test = porto_taxi.train_test_split(
    target_column="trip_id", task="HMP", test_size=0.2, n_bins=7
)

Created new train_gdf and test_gdf. Train len: 358960, test len: 89740

In [19]:

Copied!

len(train), len(test)
len(train), len(test)

Out[19]:

(358960, 89740)

In [20]:

Copied!





def visualize_h3_trajectories(
    h3_sequences, map_center=(41.14075, -8.61029), zoom_start=12
):
    """
    Visualize H3 sequences on a Folium map.

    Args:
        h3_sequences (List[List[str]]): A list of H3 sequences (trajectories).
        map_center (Tuple[float, float]): Center of the map (lat, lon).
        zoom_start (int): Initial zoom level.
    """
    m = folium.Map(location=map_center, zoom_start=zoom_start, tiles="cartodbpositron")

    colors = ["red", "blue", "green", "purple", "orange", "darkred", "lightblue"]

    for i, sequence in enumerate(h3_sequences):
        color = colors[i % len(colors)]

        for h3_id in sequence:
            boundary = h3.cell_to_boundary(
                h3_id,
            )
            folium.Polygon(
                locations=boundary, color=color, weight=2, fill=True, fill_opacity=0.3
            ).add_to(m)

    return m


h3_sequences = train["h3_sequence"].tolist()
map_ = visualize_h3_trajectories(h3_sequences[0:10])  # visualize first 10 for speed
display(map_)
def visualize_h3_trajectories(
    h3_sequences, map_center=(41.14075, -8.61029), zoom_start=12
):
    """
    Visualize H3 sequences on a Folium map.

    Args:
        h3_sequences (List[List[str]]): A list of H3 sequences (trajectories).
        map_center (Tuple[float, float]): Center of the map (lat, lon).
        zoom_start (int): Initial zoom level.
    """
    m = folium.Map(location=map_center, zoom_start=zoom_start, tiles="cartodbpositron")

    colors = ["red", "blue", "green", "purple", "orange", "darkred", "lightblue"]

    for i, sequence in enumerate(h3_sequences):
        color = colors[i % len(colors)]

        for h3_id in sequence:
            boundary = h3.cell_to_boundary(
                h3_id,
            )
            folium.Polygon(
                locations=boundary, color=color, weight=2, fill=True, fill_opacity=0.3
            ).add_to(m)

    return m


h3_sequences = train["h3_sequence"].tolist()
map_ = visualize_h3_trajectories(h3_sequences[0:10])  # visualize first 10 for speed
display(map_)

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