



European rail networks have aged past their design life. Renewal trails the work each year. The same networks now carry more traffic on essentially fixed track. Since October 2024, EU Directive 2022/2557 (CER) obliges Member States to require recurring risk assessment on critical rail operators. Why now carries the full reading.
Today, owners read critical points of infrastructure like level crossings one at a time. Studies, reviews, and mandated traffic counts exist. But coverage stays uneven, and the operational view runs as a patchwork.
What's missing is a view across the whole portfolio. One that runs systematic, homogeneous, and easier to maintain. One that adds colour to the picture by surfacing the context around each point. And one that reveals which key flows rely on which points, so the picture clarifies the dependencies.
Turning that context and that dependency view into indicators creates a shared basis. Owners can then compare points across the portfolio. They can prioritise focus areas and investment programmes against the same metrics.
SAMRoute turns scattered, uneven information into a shared, comparable picture. Teams spend less time reconciling. They spend more time deciding what to do first across a full network.
First, it expands situational context. The platform scans well beyond the usual radius, so the surrounding environment opens up consistently at scale.
Then, it clarifies the dependencies between points. The platform surfaces which key flows likely link to which points (under explicit assumptions). Portfolio discussions then rest on a more consistent basis, alongside existing studies, reviews, and local expertise.
SAMRoute pulls in general-purpose geospatial data (network topology, reference layers, context datasets). It then adds task-specific inputs, such as inventories of emitters and critical points, and customer datasets where available. The modelling choices rest on explicit assumptions that owners can review and adjust.
The platform generates origin–destination (OD) pairs and computes a primary route plus an alternative that avoids a given critical point. It keeps only the OD pairs that actually traverse the point. It then aggregates results per critical point and stores them for live use in the UI.
The chain stays traceable end-to-end, from inputs through assumptions to results. It supports sharing today. It will support machine-to-machine hooks and keep outputs current as data evolves.
This demo targets rail-infrastructure teams. It shows how the platform describes the local context around a level crossing and the dependencies of at-risk road flows on a consistent basis. It then consolidates both into a portfolio view you can sort and compare to support prioritisation.
To see it in action, request a walkthrough.

Fabrice Colas founded Oriskami SAS and runs SAMRoute. PhD in statistical learning from Leiden University, postdoctoral research at UCLA, UMC Utrecht, and Leiden University Medical Center, and a personal CIR-approved consultant agrément from the French Ministry of Research (2022, renewed 2025–2027).
Who we are carries the fuller depth panel, including selected publications, prior production-shipping work, and applied epidemio-statistics.