Accurate description of the railway network as always been a key resource to manage train circulations. To solve common problems of operation simulation requires a great amount of data of high quality.
Unfortunately, open data published by infrastructure managers in Europe is not meeting expectations: poor documentation, no information on data quality, low frequency updates, missing data, heterogeneous coverage…
In this talk, we will compare current railway network data availability and quality on OpenStreetMap and on company-owned open data platforms of several European countries. Based on this analysis, we will issue usage recommendations for open-source projects needing to use railway network description data. Then, we will share perspectives on the evolution of these data sources on the years to come.
Accurate description of the railway network as always been a key resource to manage train circulations. To solve common problems of operation simulation requires a great amount of data of high quality:
Comprehensive geometry (length, incline, curve) and topology of tracks, including switches and crossings and of passing points (stations, points of interest);
Circulation restrictions, associated to tracks (speed limit, track vacancy detection, electrification, loading gauge);
Rolling stock (power, size, weight, contents).
Historically, this data has been created and maintained by rail infrastructure managers, to construct the timetable, sell train paths to transporters and maintain the infrastructure. The strategic nature of the data has been reassessed these last years, as open data regulation evolved in Europe (public money = public data), compelling infrastructure managers to publish some of their data. Open data supports innovative product emergence inside and outside of the company, and is a requirement on open-source projects, to allow collaboration.
Unfortunately, published data is not meeting expectations: poor documentation, no information on data quality, low frequency updates, missing data, heterogeneous coverage…
Meanwhile, since 2007 on the OpenStreetMap (OSM) project, a collaborative geographic database of the world, volunteers have mapped the European railway network with great accuracy, using mainly aerial imagery and field knowledge provided by railway enthusiasts. Because of its overall quality, coverage, and easy access, projects started to use OSM data instead of company-provided open data (OpenRailwayMap, OpenRailRouting).
Using OSM data has many advantages: public and decentralized (anyone can access, at any time), constantly updated with detailed metadata, world coverage, easy updating of incorrect data, data model flexibility allowing new usages, powerful open-source tools to edit and ascertain data quality, protective license (OdBL).
But there are still many challenges to address: integrating company-sourced open data in OSM while respecting the community work (and as so, challenging the ownership of the data as a data producer), establishing public data quality reports, integrating real-time and non-geographic data (such as rolling stocks) that cannot be published on OSM, work with other users to ensure the data model and data quality suits every use, etc.
In this talk, we will compare current railway network data availability and quality on OpenStreetMap and on company-owned open data platforms of several European countries. Based on this analysis, we will issue usage recommendations for open-source projects needing to use railway network description data. Then, we will share perspectives on the evolution of these data sources on the years to come.