awesome-rerite

A curated list of awesome open datasets and benchmarks in transportation research.

Contents

Transportation Demand and Network Modeling

Network & OD Data

• Travel Forecasting, Explained — a community-maintained knowledge base for demand modeling; hat tip to Greg Macfarlane.
• Transportation Networks — curated benchmark city networks compiled by Ben Stabler.
• MATSim Scenarios — ready-to-run multi-agent transport scenarios for MATSim experiments.
• UniTA — unified traffic assignment datasets spanning 20 U.S. cities (nickname provided here until the authors coin an official one).
• WardropNet — ICLR 2025 benchmark for equilibrium-augmented traffic flow prediction.
• LargeOD — ICLR 2025 dataset for large-scale commuting OD generation.
• GNN4UE — user-equilibrium traffic flow datasets for graph neural network prediction contributed by Bahman Madadi.

Tooling & Workflows

• OSMnx — award-winning toolkit by Geoff Boeing for downloading, modeling, and analyzing street networks.
• osm2gmns — converts OpenStreetMap data into GMNS format; developed by Jiawei Lu and Xuesong (Simon) Zhou.
• DTALite — ASU Transportation AI Lab’s macroscopic/mesoscopic AMS library that natively reads GMNS.
• DTA — Dr. Ke Han’s open-source dynamic user equilibrium solver (see Networks and Spatial Economics, 2019).
• OSRM — C++14 routing engine providing lightning-fast shortest path queries.

Research Benchmarks

• WardropNet and LargeOD — paired ICLR 2025 papers outlining reproducible evaluation pipelines.
• GNN4UE — includes starter baselines for graph learning atop equilibrium traffic networks.

Discrete Choice Modeling

Benchmarks & Toolkits

• Choice-Learn — Python-first framework for training and benchmarking large-scale discrete choice models.
• DNN for Choice Modeling — code accompanying the TR-C paper on enforcing behavioral regularity with gradient penalties.
• Apollo — freeware choice-model estimation suite (Hess & Palma, 2019); shoutout to Siqi Feng for the pointer.

Datasets

• Swissmetro — canonical dataset for testing discrete choice specifications (thanks again to Siqi).

Traffic Dynamics and Control

Vehicle Trajectory Data

• pNEUMA — large-scale drone capture over Athens for microscopic trajectory analysis.
• Songdo dataset — paired traffic and vision releases documenting the Songdo smart-city corridor.
• Waymo Open Dataset — richly annotated autonomous-driving trajectories.
• I-24 MOTION — long-horizon freeway trajectories from Tennessee’s instrumented corridor.
• MiTra — drone-based all-traffic-state freeway dataset collected near Milan.
• TGSIM I-294 — third-generation simulation data for I-294 in Chicago.
• Zen Traffic Data — extensive Japanese freeway sensor archive.
• OpenACC — Adaptive Cruise Control experiments from the Joint Research Centre.
• microSIM-ACC — open code + datasets for microscopic ACC controller evaluation.

Ring-Road Experiments (The Lord of the Rings)

• Nagoya Traffic Jam Experiment — download trajectories.
• Dome Traffic Jam Experiment — circular arena trajectories.
• Arizona Ring Experiments Dataset (ARED) — autonomous-vehicle dampening experiments from Vanderbilt.

Freeway Traffic Control

• Flow-Lite — lightweight sibling of Flow without RLlib, ideal for rapid freeway controller prototyping.

Urban Traffic Control

• Intersection Zoo — ICLR 2025 benchmark for multi-agent eco-driving at signalized intersections.

Traffic Simulation

General-purpose simulators that complement the DTA-focused resources above.

• UXSim — mesoscopic simulator by Toru Seo for large-scale experiments.
• DTALite — doubles as a simulator when you need AMS-style macroscopic and mesoscopic assignment.
• sym-metanet — Python implementation of the METANET macroscopic control framework by Filippo Airaldi.

Spatio-Temporal Data and Modeling

• Spatiotemporal Data — curated catalog of datasets, models, and code for spatio-temporal learning.

Road Safety

• Surrogate Safety Measures — open implementations of plane-based SSMs from Yiru Jiao.
• SHRP2NDS-BEVTR — bird’s-eye reconstructions of SHRP2 naturalistic driving data with crash/near-crash labels (also courtesy of Yiru).

Mobility Services

• FleetPy — open-source fleet simulation framework (seven-plus reference projects and counting!).

Optimization

• RL4CO — reinforcement-learning-for-combinatorial-optimization environments collected under one documented roof.

Transportation and AI

Generative & Tutorial Resources

• DGMinTransportation — companion repo for “A Gentle Introduction and Tutorial on Deep Generative Models in Transportation Research” from UMN-Choi-Lab.

Traffic Prediction via Deep Learning

Still one of the hottest intersections between AI and transportation — here are curated entry points.

• Deep learning models for traffic prediction — Jiawei Jiang’s catalog of open-source prediction models and codebases.
• LibCity — unified benchmark covering spatio-temporal datasets, models, and evaluation tooling.

Data Standards in Transportation

• GTFS and more — MobilityData’s definitive list covering GTFS, NeTEx, and emerging mobility data specs.

Notable Figures in Transportation

• Dietrich Braess — of Braess’s Paradox fame: adding a link can slow everyone down.
• Markos Papageorgiou — start with this InControl podcast episode for a tour of modern traffic control.

Miscellaneous

• Optuna — automatic hyperparameter optimization framework (thanks to Alex Richardson for the recommendation).
• OSMnx Matplotlib animation — visualize your personal mobility over urban networks.
• Edge-bundling tool for regional mobility flow data — high-impact flow visualizations for regional planning.

How to Contribute?

If you have a dataset or benchmark to contribute, you can:

1. Open an issue

2. Submit a pull request

3. Contact me via email

---

Maintained by Junyi Ji, Views are my own. Last updated: Septempber 15, 2025.

↑ Top