PTRAIL: A Parallel TRajectory dAta preprocessIng Library

Introduction

PTRAIL is a state-of-the art Mobility Data Preprocessing Library that mainly deals with filtering data, generating features and interpolation of Trajectory Data.

The main features of PTRAIL are:

1. PTRAIL uses primarily parallel computation based on python Pandas and numpy which makes it very fast as compared to other libraries available.
2. PTRAIL harnesses the full power of the machine that it is running on by using all the cores available in the computer.
3. PTRAIL uses a customized DataFrame built on top of python pandas for representation and storage of Trajectory Data.
4. PTRAIL also provides several Temporal and spatial features which are calculated mostly using parallel computation for very fast and accurate calculations.
5. Moreover, PTRAIL also provides several filteration and outlier detection methods for cleaning and noise reduction of the Trajectory Data.
6. Apart from the features mentioned above, four different kinds of Trajectory Interpolation techniques are offered by PTRAIL which is a first in the community.

Documentation

↪ PTRAIL Documentation

Installation

1. Create Virtual Environment:

• Using Pip:
  • python3 -m venv ptr
  • source ptr/bin/activate
  • pip install PTRAIL
• Using Conda:
  • conda create -c conda-forge ptr python=3.10 rtree
  • conda activate ptr
  • pip install PTRAIL

Examples

↪ PTRAIL Examples

Miscellaneous

Citation

To cite PTRAIL in your academic work, please use the following citation:

@article{haidri2022ptrail,
  title={PTRAIL—A python package for parallel trajectory data preprocessing},
  author={Haidri, Salman and Haranwala, Yaksh J and Bogorny, Vania and Renso, Chiara and da Fonseca, Vinicius Prado and Soares, Amilcar},
  journal={SoftwareX},
  volume={19},
  pages={101176},
  year={2022},
  publisher={Elsevier}
}