OSM+: Billion-Level Open Street Map Data Processing System for City-wide Experiments

TL;DR

The paper presents OSM+, a cloud-based, structured road-network computing platform that converts OpenStreetMap data into a global, graph-structured dataset with billions of elements. It introduces a graph model and efficient spatial querying APIs implemented on MaxCompute, plus converters to support downstream analytics and model training. Three use-case benchmarks—city-scale traffic prediction, traffic policy control, and city boundary detection—demonstrate scalability and utility, complemented by novel benchmarks across 31 cities and six large-scale city scenarios. The work addresses data accessibility, consistency, and scalability gaps in open-road-network research, enabling broader, more rigorous urban studies and multimodal modeling. This dataset and tooling are poised to accelerate empirical urban science and multi-modal foundation-model research that relies on comprehensive road-network inputs.

Abstract

Road network data can provide rich information about cities and thus become the base for various urban research. However, processing large volume world-wide road network data requires intensive computing resources and the processed results might be different to be unified for testing downstream tasks. Therefore, in this paper, we process the OpenStreetMap data via a distributed computing of 5,000 cores on cloud services and release a structured world-wide 1-billion-vertex road network graph dataset with high accessibility (opensource and downloadable to the whole world) and usability (open-box graph structure and easy spatial query interface). To demonstrate how this dataset can be utilized easily, we present three illustrative use cases, including traffic prediction, city boundary detection and traffic policy control, and conduct extensive experiments for these three tasks. (1) For the well-investigated traffic prediction tasks, we release a new benchmark with 31 cities (traffic data processed and combined with our released OSM+ road network dataset), to provide much larger spatial coverage and more comprehensive evaluation of compared algorithms than the previously frequently-used datasets. This new benchmark will push the algorithms on their scalability from hundreds of road network intersections to thousands of intersections. (2) While for the more advanced traffic policy control task which requires interaction with the road network, we release a new 6 city datasets with much larger scale than the previous datasets. This brings new challenge for thousand-scale multi-agent coordination. (3) Along with the OSM+ dataset, the release of data converters facilitates the integration of multimodal spatial-temporal data for geospatial foundation model training, thereby expediting the process of uncovering compelling scientific insights. PVLDB Reference Forma

Figures (9)

• Figure 1: Comparison among four ways of querying OpenStreetMap data: website search, the Overpass API, the OSMnx package, and our proposed method (OSM+ dataset). While the first three either lack an explicit graph structure, are difficult to use, or suffer from slow query speed, our OSM+ dataset stored in a cloud database enables efficient, scalable queries over a precomputed graph-structured dataset.
• Figure 2: OSM+ database provides worldwide edges (Table split_edge) and points (Table vertex). Only basic SQL language is needed to facilitate the efficient construction of an urban graph structure on a global scale.
• Figure 3: Basic continent and category statistics of OSM+ database.
• Figure 4: OSM+ (on cloud) provides efficient and easy-to-use query service. If global data is processed on a local device (w/o OSM+), out of memory may be caused by the join operation due to the large data table (top left). In addition to this, it is cumbersome to process all the data in the region when constructing a specific region map data (top right).
• Figure 5: The framework of downstream task supporting OSM+ database.

Theorems & Definitions (1)

• definition 1: Graph-based Road Network Model