Note

This page was generated from user_guide/data_analysis/dlr-ut-dataset.pct.py.
Interactive online version: Binder badge

Using the DLRUTDataset#

This example shall give an overview of the methods and attributes that are available in the DLRUTDataset class.

Load trajectory data#

At first, we need to load the trajectory data of the dataset.

[1]:

from tasi.dlr import DLRTrajectoryDataset
from tasi.dlr.dataset import DLRUTDatasetManager, DLRUTVersion

dataset = DLRUTDatasetManager(DLRUTVersion.latest)
dataset.load()

ds = DLRTrajectoryDataset.from_csv(dataset.trajectory()[50])

[2025-04-25 12:18:50 | dataset.py:load:130] > INFO:  Checking if dataset already downloaded /tmp/DLR-Urban-Traffic-dataset_v1-2-0
[2025-04-25 12:18:50 | dataset.py:load:166] > INFO:  Dataset already available at /tmp/DLR-Urban-Traffic-dataset_v1-2-0

Attributes of the dataset#

There are several attributes available to get information about a dataset. For instance, we can get the interval of a dataset via the property

[2]:

ds.interval

[2]:

Interval(2023-09-24 12:30:00.016482+00:00, 2023-09-24 12:44:59.966482+00:00, closed='right')

or all unique timestamps of it via

[3]:

ds.timestamps

[3]:

DatetimeIndex(['2023-09-24 12:30:00.016482+00:00',
               '2023-09-24 12:30:00.066482+00:00',
               '2023-09-24 12:30:00.116482+00:00',
               '2023-09-24 12:30:00.166482+00:00',
               '2023-09-24 12:30:00.216482+00:00',
               '2023-09-24 12:30:00.266482+00:00',
               '2023-09-24 12:30:00.316482+00:00',
               '2023-09-24 12:30:00.366482+00:00',
               '2023-09-24 12:30:00.416482+00:00',
               '2023-09-24 12:30:00.466482+00:00',
               ...
               '2023-09-24 12:44:59.516482+00:00',
               '2023-09-24 12:44:59.566482+00:00',
               '2023-09-24 12:44:59.616482+00:00',
               '2023-09-24 12:44:59.666482+00:00',
               '2023-09-24 12:44:59.716482+00:00',
               '2023-09-24 12:44:59.766482+00:00',
               '2023-09-24 12:44:59.816482+00:00',
               '2023-09-24 12:44:59.866482+00:00',
               '2023-09-24 12:44:59.916482+00:00',
               '2023-09-24 12:44:59.966482+00:00'],
              dtype='datetime64[ns, UTC]', name='timestamp', length=18000, freq=None)

or the ids of all traffic participants in the dataset.

[4]:

ds.ids

[4]:

Index([1695558514712340, 1695558558975558, 1695558561769745, 1695558563216597,
       1695558565170472, 1695558571572035, 1695558572272810, 1695558573520980,
       1695558576618026, 1695558579216721,
       ...
       1695559487523035, 1695559488624461, 1695559490123751, 1695559490123748,
       1695559491223905, 1695559492674737, 1695559496120575, 1695559498274800,
       1695559499574101, 1695559499171792],
      dtype='int64', name='id', length=750)

Filtering#

If you want to look into a short sequence of the overall dataset, you can select specific rows of the overall dataset. The tasi.DLRTrajectoryDataset provides various ways for this purpose.

Time and object#

There are two variants to filter a dataset based on the information on the dataset’s index. For instance, if you want to filter the dataset by an interval, you can utilize the tasi.DLRTrajectoryDataset.during method.

[5]:

ds.during(ds.timestamps[0], ds.timestamps[10])

[5]:
acceleration center classifications dimension interpolated velocity yaw
easting magnitude northing easting northing bicycle car motorbike pedestrian truck van height length width easting magnitude northing
timestamp id
2023-09-24 12:30:00.016482+00:00 1695558514712340 1.068 1.698 1.320 604774.496 5792796.320 0.008 0.598 0.388 0.000 0.006 0.0 1.445 2.536 1.177 False 7.140 8.169 -3.970 -29.398
1695558558975558 0.014 0.016 -0.008 604740.884 5792779.240 0.000 0.838 0.163 0.000 0.000 0.0 1.261 2.669 1.259 False 0.021 0.021 0.000 7.852
1695558561769745 0.008 0.036 -0.035 604787.539 5792764.373 0.000 0.952 0.037 0.011 0.000 0.0 1.311 3.445 1.409 False -0.008 0.042 0.041 107.830
1695558563216597 -0.041 0.044 -0.014 604741.853 5792776.267 0.000 0.954 0.026 0.000 0.000 0.0 1.468 3.323 1.293 False -0.030 0.044 0.032 9.183
1695558565170472 -0.002 0.004 -0.004 604791.430 5792770.438 0.000 1.000 0.000 0.000 0.000 0.0 1.556 3.407 1.775 False -0.000 0.013 0.013 116.146
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2023-09-24 12:30:00.466482+00:00 1695558592618702 -1.060 1.065 -0.095 604725.976 5792780.341 0.000 0.904 0.095 0.000 0.001 0.0 1.496 3.387 1.496 False 3.969 4.002 0.516 7.403
1695558596566965 0.117 0.194 -0.155 604744.173 5792789.955 0.163 0.000 0.000 0.837 0.000 0.0 1.624 1.323 0.787 False 0.441 1.529 -1.464 -73.101
1695558596716951 0.054 0.243 -0.237 604742.064 5792846.878 0.000 0.933 0.062 0.000 0.000 0.0 1.368 2.848 1.394 False 0.674 2.345 -2.246 -73.240
1695558598618723 -0.116 0.137 0.072 604745.924 5792793.615 1.000 0.000 0.000 0.000 0.000 0.0 1.445 0.939 1.029 False 1.345 2.942 -2.617 -62.834
1695558598820336 0.091 0.152 -0.122 604747.604 5792840.329 0.000 0.922 0.068 0.000 0.010 0.0 1.640 2.950 1.698 False 3.606 10.793 -10.172 -70.505

172 rows × 19 columns

that returns the rows within the given interval.

Another variant to select specific rows of the datasets is by the id of a traffic participant. This might be useful if you want to take a closer look into the behavior of specific traffic participants. For instance, to filter by the second traffic participant in the dataset, we can combine the tasi.DLRTrajectoryDataset.ids attribute with the trajectory method.

[6]:

ds.trajectory(ds.ids[1])

[6]:
acceleration center classifications dimension interpolated velocity yaw
easting magnitude northing easting northing bicycle car motorbike pedestrian truck van height length width easting magnitude northing
timestamp id
2023-09-24 12:30:00.016482+00:00 1695558558975558 0.014 0.016 -0.008 604740.884 5792779.240 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 0.021 0.021 0.000 7.852
2023-09-24 12:30:00.066482+00:00 1695558558975558 0.013 0.015 -0.008 604740.884 5792779.240 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 0.023 0.023 -0.001 7.852
2023-09-24 12:30:00.116482+00:00 1695558558975558 0.012 0.014 -0.008 604740.884 5792779.240 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 0.025 0.025 -0.001 7.852
2023-09-24 12:30:00.166482+00:00 1695558558975558 0.011 0.014 -0.007 604740.884 5792779.240 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 0.028 0.028 -0.002 7.852
2023-09-24 12:30:00.216482+00:00 1695558558975558 0.010 0.013 -0.007 604740.885 5792779.240 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 0.030 0.030 -0.002 7.852
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2023-09-24 12:30:34.216482+00:00 1695558558975558 -0.242 0.304 -0.183 604783.725 5792742.854 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 3.161 10.290 -9.792 -72.008
2023-09-24 12:30:34.266482+00:00 1695558558975558 -0.242 0.304 -0.183 604783.882 5792742.364 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 3.153 10.293 -9.798 -72.049
2023-09-24 12:30:34.316482+00:00 1695558558975558 -0.242 0.304 -0.183 604784.038 5792741.873 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 3.146 10.296 -9.803 -72.084
2023-09-24 12:30:34.366482+00:00 1695558558975558 -0.242 0.304 -0.183 604784.194 5792741.382 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 3.140 10.298 -9.808 -72.114
2023-09-24 12:30:34.416482+00:00 1695558558975558 -0.242 0.304 -0.183 604784.349 5792740.890 0.0 0.838 0.163 0.0 0.0 0.0 1.261 2.669 1.259 False 3.134 10.301 -9.812 -72.141

689 rows × 19 columns

Traffic participant properties#

There are also methods available that might help to find the relevant information in the dataset. The most straight forward option is to use pandas’ capability to access specific attributes of the datasets. The available attributes on the dataset, are available via the tasi.DLRTrajectoryDataset.attribute property.

[7]:

ds.attributes

[7]:

Index(['acceleration', 'center', 'classifications', 'dimension',
       'interpolated', 'velocity', 'yaw'],
      dtype='object')

We can, for instance, access the traffic participants center position.

[8]:

ds.center

[8]:
easting northing
timestamp id
2023-09-24 12:30:00.016482+00:00 1695558514712340 604774.496 5792796.320
1695558558975558 604740.884 5792779.240
1695558561769745 604787.539 5792764.373
1695558563216597 604741.853 5792776.267
1695558565170472 604791.430 5792770.438
... ... ... ...
2023-09-24 12:44:59.966482+00:00 1695559492674737 604789.200 5792769.423
1695559496120575 604702.210 5792771.596
1695559498274800 604855.834 5792822.882
1695559499171792 604740.345 5792785.036
1695559499574101 604871.056 5792836.959

388075 rows × 2 columns

or the classification propabilities.

[9]:

ds.classifications

[9]:
bicycle car motorbike pedestrian truck van
timestamp id
2023-09-24 12:30:00.016482+00:00 1695558514712340 0.008 0.598 0.388 0.000 0.006 0.0
1695558558975558 0.000 0.838 0.163 0.000 0.000 0.0
1695558561769745 0.000 0.952 0.037 0.011 0.000 0.0
1695558563216597 0.000 0.954 0.026 0.000 0.000 0.0
1695558565170472 0.000 1.000 0.000 0.000 0.000 0.0
... ... ... ... ... ... ... ...
2023-09-24 12:44:59.966482+00:00 1695559492674737 0.000 0.979 0.021 0.000 0.000 0.0
1695559496120575 0.014 0.572 0.414 0.001 0.000 0.0
1695559498274800 0.000 0.921 0.079 0.000 0.000 0.0
1695559499171792 0.000 0.660 0.340 0.000 0.000 0.0
1695559499574101 0.022 0.890 0.089 0.000 0.000 0.0

388075 rows × 6 columns

We extended these basic capabilities with additional methods, that, for instance, allow to get the most likely class by each traffic participant’s pose

[10]:

ds.most_likely_class(by="pose")

[10]:

timestamp                         id
2023-09-24 12:30:00.016482+00:00  1695558514712340    car
                                  1695558558975558    car
                                  1695558561769745    car
                                  1695558563216597    car
                                  1695558565170472    car
                                                     ...
2023-09-24 12:44:59.966482+00:00  1695559492674737    car
                                  1695559496120575    car
                                  1695559498274800    car
                                  1695559499171792    car
                                  1695559499574101    car
Length: 388075, dtype: object

or by the overall trajectory (the default), i.e. all poses of a traffic participants.

[11]:

ds.most_likely_class(by="trajectory")

[11]:

id
1695558514712340    car
1695558558975558    car
1695558561769745    car
1695558563216597    car
1695558565170472    car
                   ...
1695559492674737    car
1695559496120575    car
1695559498274800    car
1695559499171792    car
1695559499574101    car
Name: classification, Length: 750, dtype: object

This might help to filter the dataset to select only traffic participants that are classified as a car. To archieve this, we first get the most likely class per trajectory, select the rows having the value ‘car’ and pass their index (the traffic particpant’s id) into the tasi.DLRTrajectoryDataset.trajectory method.

[12]:

classification = ds.most_likely_class(by="trajectory")

ds.trajectory(classification[classification == "car"].index)

[12]:
acceleration center classifications dimension interpolated velocity yaw
easting magnitude northing easting northing bicycle car motorbike pedestrian truck van height length width easting magnitude northing
timestamp id
2023-09-24 12:30:00.016482+00:00 1695558514712340 1.068 1.698 1.320 604774.496 5792796.320 0.008 0.598 0.388 0.000 0.006 0.0 1.445 2.536 1.177 False 7.140 8.169 -3.970 -29.398
1695558558975558 0.014 0.016 -0.008 604740.884 5792779.240 0.000 0.838 0.163 0.000 0.000 0.0 1.261 2.669 1.259 False 0.021 0.021 0.000 7.852
1695558561769745 0.008 0.036 -0.035 604787.539 5792764.373 0.000 0.952 0.037 0.011 0.000 0.0 1.311 3.445 1.409 False -0.008 0.042 0.041 107.830
1695558563216597 -0.041 0.044 -0.014 604741.853 5792776.267 0.000 0.954 0.026 0.000 0.000 0.0 1.468 3.323 1.293 False -0.030 0.044 0.032 9.183
1695558565170472 -0.002 0.004 -0.004 604791.430 5792770.438 0.000 1.000 0.000 0.000 0.000 0.0 1.556 3.407 1.775 False -0.000 0.013 0.013 116.146
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2023-09-24 12:44:59.966482+00:00 1695559492674737 0.612 2.202 -2.115 604789.200 5792769.423 0.000 0.979 0.021 0.000 0.000 0.0 1.318 3.642 1.634 False -0.813 2.602 2.471 108.191
1695559496120575 -0.850 0.857 0.111 604702.210 5792771.596 0.014 0.572 0.414 0.001 0.000 0.0 1.304 2.280 0.931 False 8.734 8.745 0.456 2.955
1695559498274800 -0.009 0.253 0.252 604855.834 5792822.882 0.000 0.921 0.079 0.000 0.000 0.0 1.385 3.026 1.385 False -13.271 14.496 -5.832 -156.208
1695559499171792 1.309 1.314 0.118 604740.345 5792785.036 0.000 0.660 0.340 0.000 0.000 0.0 1.269 2.907 1.451 False 5.195 5.274 0.908 9.847
1695559499574101 -0.422 0.423 -0.030 604871.056 5792836.959 0.022 0.890 0.089 0.000 0.000 0.0 1.389 3.028 1.593 False -15.391 16.423 -5.728 -159.568

344669 rows × 19 columns

You can achieve the same result by directly calling

[13]:

ds.cars

[13]:
acceleration center classifications dimension interpolated velocity yaw
easting magnitude northing easting northing bicycle car motorbike pedestrian truck van height length width easting magnitude northing
timestamp id
2023-09-24 12:30:00.016482+00:00 1695558514712340 1.068 1.698 1.320 604774.496 5792796.320 0.008 0.598 0.388 0.000 0.006 0.0 1.445 2.536 1.177 False 7.140 8.169 -3.970 -29.398
1695558558975558 0.014 0.016 -0.008 604740.884 5792779.240 0.000 0.838 0.163 0.000 0.000 0.0 1.261 2.669 1.259 False 0.021 0.021 0.000 7.852
1695558561769745 0.008 0.036 -0.035 604787.539 5792764.373 0.000 0.952 0.037 0.011 0.000 0.0 1.311 3.445 1.409 False -0.008 0.042 0.041 107.830
1695558563216597 -0.041 0.044 -0.014 604741.853 5792776.267 0.000 0.954 0.026 0.000 0.000 0.0 1.468 3.323 1.293 False -0.030 0.044 0.032 9.183
1695558565170472 -0.002 0.004 -0.004 604791.430 5792770.438 0.000 1.000 0.000 0.000 0.000 0.0 1.556 3.407 1.775 False -0.000 0.013 0.013 116.146
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2023-09-24 12:44:59.966482+00:00 1695559492674737 0.612 2.202 -2.115 604789.200 5792769.423 0.000 0.979 0.021 0.000 0.000 0.0 1.318 3.642 1.634 False -0.813 2.602 2.471 108.191
1695559496120575 -0.850 0.857 0.111 604702.210 5792771.596 0.014 0.572 0.414 0.001 0.000 0.0 1.304 2.280 0.931 False 8.734 8.745 0.456 2.955
1695559498274800 -0.009 0.253 0.252 604855.834 5792822.882 0.000 0.921 0.079 0.000 0.000 0.0 1.385 3.026 1.385 False -13.271 14.496 -5.832 -156.208
1695559499171792 1.309 1.314 0.118 604740.345 5792785.036 0.000 0.660 0.340 0.000 0.000 0.0 1.269 2.907 1.451 False 5.195 5.274 0.908 9.847
1695559499574101 -0.422 0.423 -0.030 604871.056 5792836.959 0.022 0.890 0.089 0.000 0.000 0.0 1.389 3.028 1.593 False -15.391 16.423 -5.728 -159.568

344669 rows × 19 columns

This works similarly for all object classes.

[14]:

ds.trucks

[14]:
acceleration center classifications dimension interpolated velocity yaw
easting magnitude northing easting northing bicycle car motorbike pedestrian truck van height length width easting magnitude northing
timestamp id
2023-09-24 12:30:00.016482+00:00 1695558573520980 0.492 1.190 -1.084 604755.520 5792806.976 0.0 0.002 0.011 0.0 0.548 0.433 3.121 4.165 1.852 False 2.015 5.623 -5.249 -69.038
2023-09-24 12:30:00.066482+00:00 1695558573520980 0.492 1.189 -1.082 604755.628 5792806.699 0.0 0.002 0.011 0.0 0.548 0.433 3.121 4.165 1.852 False 2.048 5.699 -5.319 -68.982
2023-09-24 12:30:00.116482+00:00 1695558573520980 0.492 1.187 -1.080 604755.737 5792806.418 0.0 0.002 0.011 0.0 0.548 0.433 3.121 4.165 1.852 False 2.080 5.775 -5.388 -68.927
2023-09-24 12:30:00.166482+00:00 1695558573520980 0.491 1.184 -1.077 604755.847 5792806.134 0.0 0.002 0.011 0.0 0.548 0.433 3.121 4.165 1.852 False 2.112 5.851 -5.457 -68.874
2023-09-24 12:30:00.216482+00:00 1695558573520980 0.490 1.180 -1.073 604755.960 5792805.846 0.0 0.002 0.011 0.0 0.548 0.433 3.121 4.165 1.852 False 2.145 5.927 -5.525 -68.822
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2023-09-24 12:44:59.766482+00:00 1695559485974970 0.064 0.077 0.042 604811.791 5792808.842 0.0 0.000 0.000 0.0 1.000 0.000 3.037 13.167 3.224 False 0.031 0.039 0.023 -153.082
2023-09-24 12:44:59.816482+00:00 1695559485974970 0.059 0.071 0.040 604811.792 5792808.843 0.0 0.000 0.000 0.0 1.000 0.000 3.037 13.167 3.224 False 0.030 0.038 0.023 -153.082
2023-09-24 12:44:59.866482+00:00 1695559485974970 0.054 0.066 0.038 604811.794 5792808.843 0.0 0.000 0.000 0.0 1.000 0.000 3.037 13.167 3.224 False 0.028 0.037 0.023 -153.082
2023-09-24 12:44:59.916482+00:00 1695559485974970 0.049 0.061 0.036 604811.795 5792808.843 0.0 0.000 0.000 0.0 1.000 0.000 3.037 13.167 3.224 False 0.026 0.035 0.023 -153.082
2023-09-24 12:44:59.966482+00:00 1695559485974970 0.045 0.057 0.035 604811.797 5792808.844 0.0 0.000 0.000 0.0 1.000 0.000 3.037 13.167 3.224 False 0.024 0.034 0.023 -153.082

7076 rows × 19 columns

Custom filter or transformator#

You can also build your own filter or transformator that may apply to the trajectory or pose level. For this purpose, the TrajectoryDataset.apply method can be used.

For example, let’s assume that you want to analyse the length of the different trajectories within a dataset. This may be useful for finding anomalies. In the following, we will count the number of measurements per traffic participant.

[15]:

import pandas as pd

tj_length = ds.apply(len, by="trajectory")

# create bins of width 100 measurements and count traffic participants within bins
ds_binned = pd.cut(tj_length, range(0, tj_length.max(), 100))
counts = ds_binned.value_counts().sort_index()

ax = counts.plot(kind="bar")
../../_images/user_guide_data_analysis_dlr-ut-dataset_29_0.png

If you are instead interested in the length of each trajectory in meter, we can utilize shapely. To achieve this, we convert the tasi.TrajectoryDataset to a tasi.GeoTrajectoryDataset and gain access to the shapely feature set. This enables us to use the length attribute which is the length of the geometry.

[16]:

import numpy as np

gds = ds.as_geopandas("center")
gds.set_geometry("center", inplace=True)
tj_length = gds.length

# create bins of width 100 measurements and count traffic participants within bins
ds_binned = pd.cut(tj_length, range(0, np.int32(np.round(tj_length.max())), 10))
counts = ds_binned.value_counts().sort_index()

ax = counts.plot(kind="bar")
../../_images/user_guide_data_analysis_dlr-ut-dataset_31_0.png
On this page
Show Source