Methodology — level crossings
1. The question we model
Each level crossing on a rail network sits inside a road graph. Around it sit emitters (establishments whose activity generates risk-bearing traffic) and escape targets (graph nodes toward which that traffic plausibly flows). Public data answers three questions about this scene cleanly. Where does each level crossing sit. Where does each emitter sit. What does the road graph around them look like.
One question public data does not answer.
Which emitters' nominal routes actually cross which level crossing, and therefore which emitters depend functionally on which level crossing, and how acutely.
SAMRoute imputes that missing piece. The imputation runs crossing by crossing, across the whole national portfolio, against the same reference. It replaces the qualitative presence–absence annotation that current sector practice relies on.
2. The h0/h1 framing
For every (emitter → escape target) pair around a level crossing, the road graph supports two route computations.
h0, the nominal route on the road graph between the emitter and the target.
h1, the best alternative route between the same emitter and target that detours around the level crossing.
The imputation rule runs simply, if h0 crosses the level crossing, that emitter's flow depends on it. We then record the pair (emitter, target) as an origin–destination (OD) couple attached to the crossing. The comparison between h0 and h1 quantifies how acutely the flow depends on the crossing.
SPOF (single point of failure), where the OD couple has no h1 alternative and the level crossing remains structurally required for the flow to exist on the road graph.
Prohibitive detour, where h1 exists but the cost ratio h1/h0 exceeds an operationally meaningful threshold, and the level crossing remains functionally required even though a graph alternative exists.
If h0 does not cross the level crossing, the (emitter, target) pair stays independent of it in routing terms and contributes nothing to its dependence profile.
On the current French national portfolio, a meaningful share of level crossings carry at least one OD couple in the prohibitive-detour class.
3. The pipeline in seven steps
The same pipeline runs for every crossing on the network.
Identify at-risk emitters in the surrounding zone, including SEVESO sites, logistics hubs, schools, elderly care homes, and hazardous-trade and hazardous-waste establishments.
Enumerate plausible routes from each emitter toward escape targets in the surrounding road graph.
Detect routes that cross the level crossing, which become the h0 candidates.
Search for alternatives that detour around the crossing, which become the h1 candidates.
Measure the marginal effort of the detour, including extra time, extra distance, and the ratio h1/h0.
Evaluate the relevance of h1 per OD couple, marking it as SPOF (no h1) or prohibitive detour (h1/h0 above threshold).
Qualify the level crossing on the basis of the OD set, counting how many SPOFs, how many prohibitive detours, and against which emitter mix.
Step 7 produces what the application surfaces, both at portfolio level (rank and filter the whole network) and at per-crossing level (every OD pair, with its h0, h1, ratio, and indicator flags).
4. The metric families
A vector of roughly 200 metrics describes every level crossing, grouped into five families that together characterise its nearby environment.
Population and built environment, covering population density, average building height, and settlement footprint.
At-risk emitters, drawn from the SIRENE national business registry against a curated set of activity codes from the French standard industrial classification (NAF). Five families cover the scope: transport (rail and road freight, including dangerous-goods carriers), logistics and warehousing, chemicals and fuels, hazardous-goods trade, and hazardous-waste handling.
Sensitive sites, including schools, elderly care homes, SEVESO sites, and other establishments whose presence near a level crossing matters under safety review. The list reflects working exchanges with safety auditors rather than a fixed regulatory enumeration. The public memory of accidents such as the 2017 Millas school-bus collision shapes which categories we keep in scope.
Escape targets and road-network topology, namely graph nodes that absorb redirected flow, together with the local road-graph properties that govern h0/h1 routing.
Routing economics, namely h0 and h1 metrics (time, distance), the marginal detour ratio, and the SPOF and prohibitive-detour indicators.
The selection of NAF codes proceeds by substitution. A standardised "at-risk emitter" registry does not yet exist. NAF activity codes serve as a surrogate indicator. We publish the curated subset as part of the model documentation.
5. Coverage, freshness, and known limits
The current production scope covers 100% of the active level crossings on the French national rail network, about 12,000 crossings.
The data underpinning the model lives as a snapshot frozen at the time of analysis. Continuous refresh against live data feeds sits as the next operational step on the production roadmap. Open-data availability outside France varies. That variation paces the geographic extension to other networks.
6. Reproducibility
Every per-asset indicator unfolds back to the foundational metrics, the statistical estimates composed from them, and the reference materials that produced them. The same inputs produce the same outputs. We document source layers, version stamps, and the curated NAF subset per release. The dataset catalogue lives at api.samroute.com/api/public/data-catalog.
7. Reach us
For methodology questions, technical due diligence, or model-card review, write to contact@samroute.com.
8. Annex — See the demo
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.