This is a tool for generating timeless traffic-like graphs suitable for congestion prediction.

Make graph dataset with tools found in RoadGraphGeneratorTOI.py

Make a loader using loader.py

Make manipulations on an existing graph with tools found in graphmanipulator.py

A minimal tutorial is found in Notebook_for_TØIRoads.ipynb

An algorithm for the main graph generation is found in the paper TØIRoads: A Road Data Model Generation Tool by Grunde Wesenberg and Ana Ozaki. This corresponds to lines 138-161 in RoadGraphGeneratorTOI.py, GraphMaker.generate_graphs(). TorchDatasetMaker.generate_graphs_with_congestion collects the output network into a torch dataset.

Import RoadGraphGeneratorTOI as RGG RGG.TorchDataSetMaker.generate_torch_dataset_with_congestion makes a list of pytorch geometric Data objects, each containing a graph dataset as per the graph generation specifications. It uses RGG.GraphMaker.generate_graphs to generate each single graph.

This code is associated with the paper "TØIRoads: A Road Data Model Generation Tool", accepted for publication in TGDK, Volume 2, Issue 2 (2024 or 2025). DOI or publication details will be added when available.