Example 02 - Introduction to Trajectory Data

This example will introduce:

• Load the map matched trajectory data with mtldp.utils.data_io

• Basic classes for the trajectory data

After the trajectory data is matched on the map, you can load the trajectory by the data platform (mtldp) package.

1 Load trajectory data

1.1 Load trajectory and matched points from CSV file

from mtldp.utils import show_attributes_and_methods
from mtldp.utils.data_io import build_trajectory_dict_from_csv
import time

s = time.time()
trajs_dict = build_trajectory_dict_from_csv(trajectories_filename='data/2021-09-27_trajs.csv',points_filename='data/2021-09-27_points.csv')
time.time() - s

35.63425016403198

for traj in trajs_dict:
    break

len(traj.df)

10

show_attributes_and_methods(trajs_dict)

attributes: {'dict'}

methods: {'get_points_df', 'append', 'del_trajectory', 'get_trajs_df', 'groupby', 'add_trajectory'}

show_attributes_and_methods(traj)

attributes: {'link_id', 'intervals', 'valid', 'stop_nums', 'spill_over', 'trip_id', 'date', 'stop_delay', 'queue_ratio', 'movement_id', 'control_delay', 'traj_id', 'arrival_time', 'service_level', 'veh_length', 'veh_width', 'cycle_failure', 'eta', 'tod', 'stop_details', 'df', 'junction_id', 'text', 'traj_attributes', 'free_v', 'timestamp', 'queue_dis', 'phase_id', 'date_time', 'total_time', 'pts_num', 'avg_speed', 'total_distance'}

methods: {'check_point_attribute', 'get_traj_df', 'get_point_attributes', 'get_points_df', 'get_time_range', 'add_traj_attributes', 'set_point_attributes', 'drop_duplicates', 'update_trajs_attributes_from_df'}

2 Trajectory data structure

There are two data structures to store the trajectory data TrajectoryDict, and Trajectory.

2.1 TrajectoryDict

TrajectoryDict is new class that mainly contains a dictionary of all the Trajectory objects in the .dict member. The key of .dict is trip_id, the value is Trajectory object.

len(trajs_dict.dict)

637779

2.3 Trajectory

Trajectory contains all attributes of a trajectory. You can fetch a trajectory from a TrajectoryDict object.

traj = trajs_dict.dict["006eb1ddec08e2ffac0a01c7f11c487316a8a44f-1632745748"]

traj.df[0:3]

	trip_id	traj_id	junction_id	link_id	segment_id	movement_id	phase_id	timestamp	date	date_time	tod	latitude	longitude	distance	arrival_time	speed	error	heading	elevation
3610	006eb1ddec08e2ffac0a01c7f11c487316a8a44f	006eb1ddec08e2ffac0a01c7f11c487316a8a44f-16327...	62127887	62058105_62127887	7558579851	62058105_62127887_61977976	2	1.632746e+09	2021-09-27	08:28	8.472778	42.547322	-83.192781	-221.689549	1.632746e+09	17.597236	1.138452	182.0	194.0
3611	006eb1ddec08e2ffac0a01c7f11c487316a8a44f	006eb1ddec08e2ffac0a01c7f11c487316a8a44f-16327...	62127887	62058105_62127887	7558579851	62058105_62127887_61977976	2	1.632746e+09	2021-09-27	08:28	8.473611	42.547311	-83.193415	-169.723550	1.632746e+09	16.319458	2.554720	182.0	195.0
3612	006eb1ddec08e2ffac0a01c7f11c487316a8a44f	006eb1ddec08e2ffac0a01c7f11c487316a8a44f-16327...	62127887	62058105_62127887	7558579851	62058105_62127887_61977976	2	1.632746e+09	2021-09-27	08:28	8.474722	42.547302	-83.193949	-125.954757	1.632746e+09	13.438900	4.274432	182.0	194.0

from mtldp.utils.time import timestamp_to_date_time_and_tod

timestamp_to_date_time_and_tod(1632745746.6686296)

('2021-09-27', '12:29', 12.483333333333333)

times, lats, lons, errors = traj.get_point_attributes("timestamp", "latitude", "longitude", "error")

times[:3], lats[:3]

([1632745702.624, 1632745705.621, 1632745709.119],
 [42.547322, 42.547311, 42.547302])

# Plot the shape of the trajectory

import matplotlib.pyplot as plt

plt.plot(lons, lats, "b.-")
plt.show()

You can plot the time-space diagram of a trajectory at a movement.