Bytero Road™
Motor accident risk at the level of a single address.
Bytero Road scores the road environment a vehicle is garaged in — the geographic factor behind motor territory rating. Built from police crash registers, the state traffic census and a per-road accident model. The output is a territory factor against the national average, the dangerous corridors near the address, and map layers where the prediction can be checked against recorded crashes.
Motor accident risk
The geography of motor risk: from a pricing factor to validation against real recorded crashes.
A multiple of risk per driven kilometre versus the national average: the geographic term for an insurer’s frequency model.
A 0 to 100 road-environment score around the garaging address: road classes, junctions, curves and level crossings.
Every road rated on its own: the model scores crash risk per kilometre driven across the entire network.
Corridors where crashes physically pile up: road danger times real traffic intensity from the census.
912 thousand geocoded police crashes as the training base and the ground-truth map for validation.
The predicted crash rate on every road, verified by overlap with recorded crashes.
How Bytero Road computes risk
- 1
Road network
Classes, junctions, curves, crossings and grade from map data for the whole country.
- 2
Crash model
A model trained on 912k geocoded crashes, normalized by traffic intensity.
- 3
Factor per address
The typical driving surroundings of an address roll up into a territory factor for pricing.