Tutorials on Deep Generative Models in Transportation Research

This repository contains tutorials presenting the use of deep generative models in transportation research, as presented in the paper "A Gentle Introduction and Tutorial on Deep Generative Models in Transportation Research". The examples focus on two key types of data commonly used in transportation research: 1) travel survey data and 2) highway speed contour data (Time-Space diagrams). Each tutorial is designed to guide transportation researchers in applying deep generative models to their projects.

All code is provided in Python Jupyter Notebook format and can be executed in Google Colab for easy access and experimentation.

Citations

If you use this code for your research, please cite our paper.

@article{choi2024DGMTR,
  title={A Gentle Introduction and Tutorial on Deep Generative Models in Transportation Research},
  author={Choi, Seongjin and Jin, Zhixiong and Ham, Seungwoo and Kim, Jiwon and Sun, Lijun},
  journal={arXiv preprint arXiv:2410.07066},
  year={2024}
}

Travel Survey Data Generation Example (TravelSurvey folder; Section 4.1 in the paper)

In this example, we present the generation of synthetic travel survey data using deep generative models.

preprocessed data : https://drive.google.com/drive/folders/1zi3KmBC1_XSK6neEIRBgf7vJlGBNeh8C?usp=sharing

Time Space Diagram Generation Example (TimeSpaceDiagram folder; Section 4.2 in the paper)

In this example, we present the generation of synthetic time-space diagrams using deep generative models.

preprocessed data : https://drive.google.com/drive/folders/1EndLnkkhRBFxxMSqKZFYZzHHzKJ_SZaM?usp=sharing

pretrained model : https://drive.google.com/drive/folders/1wHLzQ3ns_wtk6EgnlQIZ-TW6fNoHbSNg?usp=sharing