Trajectory Datasets and Full version of Our VLDB 2017 paper
1. Full Version with More Proof and Experimental Details [ pdf ]
2. Trajectory Datasets (TRAJ)
Taxi trajectory data , referred to as Taxi, is the GPS trajectories collected by taxies equipped with GPS sensors
in Beijing during a period from Nov 1, 2010 to Nov 30, 2010.
The sampling rate was one point per minute. Currently, a small part of Taxi (100 trajectories) are available, and a trajectory has 39,100 data points on average.
Truck trajectory data , referred to as Truck, is the GPS trajectories collected by trucks equipped with
GPS sensors in China during a period from Aug. 2015 toOct. 2015. The sampling rate varied from 1s to 60s.
Currently, a small part of Truck (100 trajectories) are available, and they mostly have around 50 to 90 thousand data points.
GeoLife trajectory data , refered to as GeoLife, is the GPS trajectories collected in GeoLife project by 182 users in
a period from Apr. 2007 to Oct. 2011. These trajectories have a variety of sampling rates, among which 91% are logged in each 1-5 seconds or each 5-10 meters per point.
The longest trajectory has 2,156,994 points. The dataset can be found at: https://www.microsoft.com/en-us/download/details.aspx?id=52367
Each taxi or truck trajectory is saved as a separate text file. In these files, a data point is saved as a line with the format:
Here, (1) id, long, is the id of the moving object,
(2) timestamp, long, is the timestamp of sampling,
(3) latitude, double, is thelatitude of the moving object at the sampling time,
(4) longitude,double, is thelongitude of the moving object at the sampling time,
(5) speed, double, is the speed (km/h) of the moving object,
(6) direction, double, is the included angle between the heading direction of the vehicle and due north.
3. Citing TRAJ
Please fell free to use our datasets, and we encourage you to cite our VLDB 2017 paper if you have used them in your work.
Last updated on Feb 28, 2017.